The Computer Incident Response Center Luxembourg (CIRCL) is a government-driven initiative designed to provide a systematic response facility to computer security threats and incidents.
- Website
- https://www.circl.lu
- info@circl.lu
- GNA identifier
- GNA-1 GCVE registry Recent publications
Recent vulnerabilities
355 GCVE records assigned by this organization as GNA-1GCVE-1-2026-20234 (CVE-2026-95659)
Vulnerability from gna-1 – Published: 2026-09-22 12:40 – Updated: 2026-09-22 12:41| URL | Tags |
|---|---|
| https://github.com/MISP/MISP/commit/23b879073 | patch |
qwen3.8:27b
advisory
bcp-05-x-01bcp-05-x-02
Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.
| Model | Source | Identifier |
|---|---|---|
| qwen3.8:27b | ollama | qwen3.8:27b |
- Generator
-
patch2vuln.pyon 2026-09-22 12:33 - Model
qwen3.8:27b- Input
-
https://github.com/MISP/MISP/commit/23b879073.patch
4476bdd487c3… - Confidence
- high
| Commit | Subject | Patch SHA-256 |
|---|---|---|
23b879073eef
|
fix: [security] Reflected XSS via the object type in the | 4476bdd487c3… |
Fix summary
The fix introduces two layers of defense. First, the viewForObject action now validates the object_type parameter against the AnalystData::valid_targets whitelist and rejects any value not in that list with a NotFoundException, preventing arbitrary strings from reaching the view layer. Second, the two output sinks in the Overmind thread.ctp template now apply the h() HTML-encoding function to the object type before interpolation into translated strings, ensuring that even if a valid type were to contain special characters, it would be rendered as inert text rather than executable markup.
Patch summary
In AnalystDataController.php, the patch adds an import of the AnalystData model and inserts a guard clause at the top of viewForObject that checks the $object_type parameter against AnalystData::valid_targets using a strict in_array comparison; if the value is not in the list, a NotFoundException is thrown. In thread.ctp, two occurrences of strtolower($objectType) passed directly into __() translation calls are wrapped with h() to HTML-encode the value before it is embedded in the rendered HTML output.
CVSS rationale
AV:N: The vulnerability is exploitable over the network via a crafted URL. AC:L: No race conditions or special environment conditions are required; the attacker simply crafts a URL with a malicious object_type value. AT:N: No manipulation of the attack target is needed. PR:L: The victim must be authenticated to MISP to reach the AnalystData view. UI:A: The victim must actively navigate to the attacker-supplied URL (e.g., click a link), which constitutes active user interaction. VC:N, VI:N, VA:N: The vulnerable MISP server itself is not directly compromised; the impact is on the victim's browser session. SC:L: The attacker can read data available in the victim's browser context (cookies, session tokens, page content) but impact is limited to the current page. SI:L: The attacker can modify the rendered page or submit forms on behalf of the victim within the session. SA:N: No broader security impact on other systems is expected from this reflected XSS.
Weakness rationale
- CWE-79 The object_type URL parameter is reflected into HTML output without encoding, allowing script injection. This is a textbook reflected XSS.
- CWE-116 The view template interpolated the object type into translated strings without applying h() HTML encoding, which is the direct encoding failure that permits the XSS.
- CWE-20 The controller accepted any arbitrary string as the object_type parameter without validating it against the known set of valid analyst-data target types, allowing attacker-controlled data to reach the output layer.
Attack pattern rationale
- CAPEC-127 The vulnerability is a reflected XSS: attacker-controlled input from the URL is reflected into the HTML response without encoding. CAPEC-127 is the most specific and directly applicable attack pattern. No uncertainty in this mapping.
- CAPEC-61 CAPEC-61 is the broader parent pattern for XSS attacks. It is included as a secondary mapping to cover the general class of the vulnerability. CAPEC-127 is the more precise match; CAPEC-61 is listed for completeness.
Assumptions to verify
- The affected version range is inferred from the tag_version_boundary metadata (v2.5.47 with 58 commits after the fix), suggesting the fix is included in v2.5.47. The exact first affected version is not specified in the patch.
- PR:L assumes the AnalystData viewForObject action requires authentication, which is consistent with MISP's typical access model but not explicitly confirmed in the patch.
- UI:A assumes the victim must actively navigate to a crafted URL (e.g., click a link). If the malicious URL could be loaded in an iframe or auto-redirected, UI:P might be more appropriate, but the evidence supports active navigation.
- SC:L and SI:L reflect the typical impact of reflected XSS limited to the current page context; a more severe impact (SC:H, SI:H) is possible if the XSS can exfiltrate full session credentials or perform privileged actions, but the patch evidence does not confirm the scope of available data.
- The CAPEC-61 mapping is a broader parent pattern included for completeness; CAPEC-127 is the primary and most precise match.
Model comparison
Selected qwen3.8:27b
by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.
| Model | Score | Agreement | Confidence | Assumptions |
|---|---|---|---|---|
qwen3.8:27b |
6 | 9 | high | 5 |
{
"containers": {
"cna": {
"affected": [
{
"defaultStatus": "unaffected",
"modules": [
"AnalystDataController",
"Overmind theme AnalystData thread element"
],
"product": "MISP",
"programFiles": [
"app/Controller/AnalystDataController.php",
"app/View/Themed/Overmind/Elements/AnalystData/thread.ctp"
],
"repo": "https://github.com/MISP/MISP",
"vendor": "MISP",
"versions": [
{
"lessThan": "2.5.47",
"status": "affected",
"version": "0",
"versionType": "semver"
}
]
}
],
"credits": [
{
"lang": "en",
"type": "reporter",
"value": "Jeroen Pinoy"
},
{
"lang": "en",
"type": "remediation developer",
"value": "iglocska"
},
{
"lang": "en",
"type": "remediation developer",
"value": "Claude Fable 5.1"
}
],
"descriptions": [
{
"lang": "en",
"supportingMedia": [
{
"base64": false,
"type": "text/html",
"value": "\u003cp\u003eMISP contains a reflected cross-site scripting (XSS) vulnerability in the AnalystDataController::viewForObject action. The method accepted a parent object type parameter from the URL without validation and passed it to the Overmind-themed AnalystData thread view element, where it was interpolated into two translated strings and rendered into the HTML response without output encoding. An authenticated attacker who can induce a victim to navigate to a crafted URL can inject arbitrary JavaScript that executes in the victim\u0027s browser within the MISP application context. This may allow the attacker to read session data, manipulate the page, or perform actions on behalf of the victim.\u0026nbsp;\u003c/p\u003e\u003cp\u003eThe vulnerability requires the victim to be authenticated to MISP and to actively visit the attacker-supplied URL. The affected component is the AnalystData controller and the Overmind theme\u0027s AnalystData thread element.\u003c/p\u003e\u003cp\u003eVersion affected: \u0026lt;2.5.47\u003c/p\u003e"
}
],
"value": "MISP contains a reflected cross-site scripting (XSS) vulnerability in the AnalystDataController::viewForObject action. The method accepted a parent object type parameter from the URL without validation and passed it to the Overmind-themed AnalystData thread view element, where it was interpolated into two translated strings and rendered into the HTML response without output encoding. An authenticated attacker who can induce a victim to navigate to a crafted URL can inject arbitrary JavaScript that executes in the victim\u0027s browser within the MISP application context. This may allow the attacker to read session data, manipulate the page, or perform actions on behalf of the victim.\u00a0\n\nThe vulnerability requires the victim to be authenticated to MISP and to actively visit the attacker-supplied URL. The affected component is the AnalystData controller and the Overmind theme\u0027s AnalystData thread element.\n\nVersion affected: \u003c2.5.47"
}
],
"impacts": [
{
"capecId": "CAPEC-127",
"descriptions": [
{
"lang": "en",
"value": "CAPEC-127 Cross Site Scripting (Reflected)"
}
]
},
{
"capecId": "CAPEC-61",
"descriptions": [
{
"lang": "en",
"value": "CAPEC-61 Cross Site Scripting"
}
]
}
],
"metrics": [
{
"cvssV4_0": {
"Automatable": "NOT_DEFINED",
"Recovery": "NOT_DEFINED",
"Safety": "NOT_DEFINED",
"attackComplexity": "LOW",
"attackRequirements": "NONE",
"attackVector": "NETWORK",
"baseScore": 4.8,
"baseSeverity": "MEDIUM",
"privilegesRequired": "LOW",
"providerUrgency": "NOT_DEFINED",
"subAvailabilityImpact": "NONE",
"subConfidentialityImpact": "LOW",
"subIntegrityImpact": "LOW",
"userInteraction": "ACTIVE",
"valueDensity": "NOT_DEFINED",
"vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:A/VC:N/VI:N/VA:N/SC:L/SI:L/SA:N",
"version": "4.0",
"vulnAvailabilityImpact": "NONE",
"vulnConfidentialityImpact": "NONE",
"vulnIntegrityImpact": "NONE",
"vulnerabilityResponseEffort": "NOT_DEFINED"
},
"format": "CVSS",
"scenarios": [
{
"lang": "en",
"value": "GENERAL"
}
]
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-79",
"description": "CWE-79 Improper Neutralization of Input During Web Page Generation (\u0027Cross-site Scripting\u0027)",
"lang": "en",
"type": "CWE"
}
]
},
{
"descriptions": [
{
"cweId": "CWE-116",
"description": "CWE-116 Improper Encoding or Escaping of Output",
"lang": "en",
"type": "CWE"
}
]
},
{
"descriptions": [
{
"cweId": "CWE-20",
"description": "CWE-20 Improper Input Validation",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"orgId": "00000000-0000-4000-9000-000000000000"
},
"references": [
{
"name": "Security patch",
"tags": [
"patch"
],
"url": "https://github.com/MISP/MISP/commit/23b879073"
}
],
"solutions": [
{
"lang": "en",
"supportingMedia": [
{
"base64": false,
"type": "text/html",
"value": "\u003cp\u003eThe fix introduces two layers of defense. First, the viewForObject action now validates the object_type parameter against the AnalystData::valid_targets whitelist and rejects any value not in that list with a NotFoundException, preventing arbitrary strings from reaching the view layer. Second, the two output sinks in the Overmind thread.ctp template now apply the h() HTML-encoding function to the object type before interpolation into translated strings, ensuring that even if a valid type were to contain special characters, it would be rendered as inert text rather than executable markup.\u003c/p\u003e"
}
],
"value": "The fix introduces two layers of defense. First, the viewForObject action now validates the object_type parameter against the AnalystData::valid_targets whitelist and rejects any value not in that list with a NotFoundException, preventing arbitrary strings from reaching the view layer. Second, the two output sinks in the Overmind thread.ctp template now apply the h() HTML-encoding function to the object type before interpolation into translated strings, ensuring that even if a valid type were to contain special characters, it would be rendered as inert text rather than executable markup."
}
],
"source": {
"discovery": "UNKNOWN"
},
"title": "MISP Reflected XSS via Unvalidated Object Type in AnalystData Overmind Thread",
"x_gcve": [
{
"extensions": {
"bcp-05-x-01": {
"ai_annotations": [
{
"ai_level": "generated",
"description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
"gna_source": 1,
"models": [
{
"gna_source": 1,
"identifier": "qwen3.8:27b",
"name": "qwen3.8:27b",
"source": "ollama"
}
],
"review_status": "full",
"scope": "record",
"tags": [
"ai-computer-assisted:llm-generated",
"ai-computer-assisted:classification"
]
}
]
},
"bcp-05-x-02": {
"x_patch2vuln": {
"assumptions": [
"The affected version range is inferred from the tag_version_boundary metadata (v2.5.47 with 58 commits after the fix), suggesting the fix is included in v2.5.47. The exact first affected version is not specified in the patch.",
"PR:L assumes the AnalystData viewForObject action requires authentication, which is consistent with MISP\u0027s typical access model but not explicitly confirmed in the patch.",
"UI:A assumes the victim must actively navigate to a crafted URL (e.g., click a link). If the malicious URL could be loaded in an iframe or auto-redirected, UI:P might be more appropriate, but the evidence supports active navigation.",
"SC:L and SI:L reflect the typical impact of reflected XSS limited to the current page context; a more severe impact (SC:H, SI:H) is possible if the XSS can exfiltrate full session credentials or perform privileged actions, but the patch evidence does not confirm the scope of available data.",
"The CAPEC-61 mapping is a broader parent pattern included for completeness; CAPEC-127 is the primary and most precise match."
],
"capecRationale": [
{
"capecId": "CAPEC-127",
"rationale": "The vulnerability is a reflected XSS: attacker-controlled input from the URL is reflected into the HTML response without encoding. CAPEC-127 is the most specific and directly applicable attack pattern. No uncertainty in this mapping."
},
{
"capecId": "CAPEC-61",
"rationale": "CAPEC-61 is the broader parent pattern for XSS attacks. It is included as a secondary mapping to cover the general class of the vulnerability. CAPEC-127 is the more precise match; CAPEC-61 is listed for completeness."
}
],
"commit": "23b879073eefce37e1065295d180c4bea7cef95f",
"confidence": "high",
"credits": [
{
"lang": "en",
"type": "reporter",
"value": "Jeroen Pinoy"
},
{
"lang": "en",
"type": "remediation developer",
"value": "iglocska"
},
{
"lang": "en",
"type": "remediation developer",
"value": "Claude Fable 5.1"
}
],
"cvssRationale": "AV:N: The vulnerability is exploitable over the network via a crafted URL. AC:L: No race conditions or special environment conditions are required; the attacker simply crafts a URL with a malicious object_type value. AT:N: No manipulation of the attack target is needed. PR:L: The victim must be authenticated to MISP to reach the AnalystData view. UI:A: The victim must actively navigate to the attacker-supplied URL (e.g., click a link), which constitutes active user interaction. VC:N, VI:N, VA:N: The vulnerable MISP server itself is not directly compromised; the impact is on the victim\u0027s browser session. SC:L: The attacker can read data available in the victim\u0027s browser context (cookies, session tokens, page content) but impact is limited to the current page. SI:L: The attacker can modify the rendered page or submit forms on behalf of the victim within the session. SA:N: No broader security impact on other systems is expected from this reflected XSS.",
"fixSummary": "The fix introduces two layers of defense. First, the viewForObject action now validates the object_type parameter against the AnalystData::valid_targets whitelist and rejects any value not in that list with a NotFoundException, preventing arbitrary strings from reaching the view layer. Second, the two output sinks in the Overmind thread.ctp template now apply the h() HTML-encoding function to the object type before interpolation into translated strings, ensuring that even if a valid type were to contain special characters, it would be rendered as inert text rather than executable markup.",
"generatedAt": "2026-09-22T12:33:44.255419Z",
"generator": "patch2vuln.py",
"model": "qwen3.8:27b",
"modelComparison": {
"rankings": [
{
"agreementScore": 9,
"assumptionCount": 5,
"confidence": "high",
"model": "qwen3.8:27b",
"score": 6
}
],
"selectedModel": "qwen3.8:27b",
"selectionMethod": "deterministic-consensus-v1",
"selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
},
"patchSha256": "4476bdd487c3b17c3cffff36a76763c107e8bf284095d61bd6860865ce8912e0",
"patchSummary": "In AnalystDataController.php, the patch adds an import of the AnalystData model and inserts a guard clause at the top of viewForObject that checks the $object_type parameter against AnalystData::valid_targets using a strict in_array comparison; if the value is not in the list, a NotFoundException is thrown. In thread.ctp, two occurrences of strtolower($objectType) passed directly into __() translation calls are wrapped with h() to HTML-encode the value before it is embedded in the rendered HTML output.",
"patchTruncated": false,
"patches": [
{
"commit": "23b879073eefce37e1065295d180c4bea7cef95f",
"patchSha256": "4476bdd487c3b17c3cffff36a76763c107e8bf284095d61bd6860865ce8912e0",
"source": "https://github.com/MISP/MISP/commit/23b879073.patch",
"sourceUrl": "https://github.com/MISP/MISP/commit/23b879073.patch",
"subject": "fix: [security] Reflected XSS via the object type in the"
}
],
"source": "https://github.com/MISP/MISP/commit/23b879073.patch",
"subject": "fix: [security] Reflected XSS via the object type in the",
"tagVersionBoundary": {
"commits_after_fix": 58,
"repository": "https://github.com/MISP/MISP",
"tag": "v2.5.47",
"version": "2.5.47",
"version_type": "semver"
},
"weaknessRationale": [
{
"cweId": "CWE-79",
"rationale": "The object_type URL parameter is reflected into HTML output without encoding, allowing script injection. This is a textbook reflected XSS."
},
{
"cweId": "CWE-116",
"rationale": "The view template interpolated the object type into translated strings without applying h() HTML encoding, which is the direct encoding failure that permits the XSS."
},
{
"cweId": "CWE-20",
"rationale": "The controller accepted any arbitrary string as the object_type parameter without validating it against the known set of valid analyst-data target types, allowing attacker-controlled data to reach the output layer."
}
]
}
}
},
"recordType": "advisory",
"vulnId": "GCVE-1-2026-20234"
}
],
"x_generator": {
"engine": "Vulnogram 0.2.0"
}
}
},
"cveMetadata": {
"assignerOrgId": "00000000-0000-4000-9000-000000000000",
"cveId": "CVE-2026-95659",
"datePublished": "2026-09-22T12:40:00.000Z",
"dateReserved": "2026-09-22T12:40:00.000Z",
"dateUpdated": "2026-09-22T12:41:23.369639Z",
"requesterUserId": "00000000-0000-4000-9000-000000000000",
"serial": 1,
"state": "PUBLISHED",
"vulnId": "GCVE-1-2026-20234"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.1"
}
GCVE-1-2026-20261 (CVE-2026-95658)
Vulnerability from gna-1 – Published: 2026-09-22 12:31 – Updated: 2026-09-22 12:31- CWE-352 - Cross-Site Request Forgery (CSRF)
| URL | Tags |
|---|---|
| https://github.com/MISP/MISP/commit/f05e0d1c2 | patch |
qwen3.8:27b
advisory
bcp-05-x-01bcp-05-x-02
Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.
| Model | Source | Identifier |
|---|---|---|
| qwen3.8:27b | ollama | qwen3.8:27b |
- Generator
-
patch2vuln.pyon 2026-09-22 12:15 - Model
qwen3.8:27b- Input
-
https://github.com/MISP/MISP/commit/f05e0d1c2.patch
3f8a00ca2976… - Confidence
- high
| Commit | Subject | Patch SHA-256 |
|---|---|---|
f05e0d1c263b
|
fix: [security] Keep the CSRF check on stateless workflow | 3f8a00ca2976… |
Fix summary
The moduleStatelessExecution action is removed from the unlockedActions list, restoring the CSRF token check. A new mechanism (_csrfTokenHeaderOnly) is introduced that enforces the CSRF token while exempting only the field hash, which the legitimate caller (the module dialog) cannot produce because it posts a hand-built object. The module dialog's AJAX request is updated to send the CSRF token in the X-CSRF-Token header, satisfying the token check without requiring a field hash.
Patch summary
In WorkflowsController::beforeFilter(), the line adding 'moduleStatelessExecution' to $this->Security->unlockedActions is replaced with a call to $this->_csrfTokenHeaderOnly(['moduleStatelessExecution']), which enforces CSRF token validation while waiving only the field hash requirement. In execute_module.ctp, the jQuery $.ajax() call is modified to include the header 'X-CSRF-Token' populated from window.csrfToken, so the legitimate client-side caller passes the token check via the header rather than a form field.
CVSS rationale
AV:N: The attack is delivered over the network via a malicious web page. AC:L: Crafting a cross-site form post is straightforward. AT:N: No special conditions or timing are required. PR:N: The attacker requires no account or privileges on the MISP instance. UI:A: The victim (a site admin) must actively visit the attacker-controlled page or click a link that triggers the form submission. VC:N: No confidentiality impact is evident. VI:H: The attacker can write arbitrary blocklist and warninglist entries, significantly compromising the integrity of MISP's security data. VA:N: No availability impact. SC/SI/SA:N: No evidence of impact on other systems beyond MISP itself.
Weakness rationale
- CWE-352 The vulnerable action was placed in unlockedActions, which disabled the CSRF token check. This allowed a cross-site form post to trigger a state-changing action (workflow module execution) on behalf of an authenticated administrator, which is the textbook definition of CSRF.
Attack pattern rationale
- CAPEC-124 The attack pattern described in the commit message is a cross-site form post that causes an authenticated site admin to execute a workflow module with attacker-chosen parameters. This maps directly to CAPEC-124, which describes an attacker exploiting the trust a web application places in a user's browser session to perform unauthorized actions. The mapping is confident because the commit message explicitly describes a cross-site form post bypassing CSRF protection.
Assumptions to verify
- The affected version range is inferred from the tag_version_boundary metadata (v2.5.47, 6 commits after fix), implying the fix is included in v2.5.47 and prior versions are affected. The exact introduction version of the vulnerable code is unknown.
- UI:A is used because the CSRF attack requires the victim to actively navigate to the attacker's page or click a link; a purely passive load (e.g., embedded iframe auto-submit) would map to UI:P, but the commit message describes a 'cross-site form post' which typically requires at least a page visit.
- SI:N is chosen because the evidence only demonstrates integrity impact on MISP's own blocklist and warninglist data; no evidence supports impact on downstream or external systems.
- The CAPEC-124 mapping is confident based on the explicit description of a cross-site form post in the commit message.
- The commit message states the issue was 'found during the internal review, not externally reported,' but does not explicitly name a finder; therefore no finder credit is assigned.
- The _csrfTokenHeaderOnly method is referenced in the patch but its implementation is not shown; it is assumed to enforce CSRF token validation while exempting the field hash, as described in the commit message.
Model comparison
Selected qwen3.8:27b
by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.
| Model | Score | Agreement | Confidence | Assumptions |
|---|---|---|---|---|
qwen3.8:27b |
5 | 9 | high | 6 |
{
"containers": {
"cna": {
"affected": [
{
"modules": [
"WorkflowsController",
"Workflows module execution"
],
"product": "MISP",
"programFiles": [
"app/Controller/WorkflowsController.php",
"app/View/Elements/Workflows/execute_module.ctp"
],
"repo": "https://github.com/MISP/MISP",
"vendor": "MISP",
"versions": [
{
"lessThan": "2.5.47",
"status": "affected",
"version": "0",
"versionType": "semver"
}
]
}
],
"credits": [
{
"lang": "en",
"type": "remediation developer",
"value": "iglocska"
},
{
"lang": "en",
"type": "remediation developer",
"value": "Claude Opus 5 (1M context)"
}
],
"descriptions": [
{
"lang": "en",
"supportingMedia": [
{
"base64": false,
"type": "text/html",
"value": "\u003cp\u003eMISP\u0027s WorkflowsController exposed the moduleStatelessExecution action in the Security component\u0027s unlockedActions list. In CakePHP, listing an action in unlockedActions disables both the CSRF token check and the field hash validation for that action. Because moduleStatelessExecution executes a workflow module\u0027s exec() method with caller-supplied input and parameters, the absence of CSRF protection allowed an attacker to craft a cross-site form post (or equivalent cross-origin request) that, when submitted by an authenticated site administrator, would cause the administrator\u0027s browser to invoke the action on the MISP instance.\u0026nbsp;\u003c/p\u003e\u003cp\u003eThe attacker could select any workflow module to execute, including action modules that write blocklist and warninglist entries, and supply arbitrary input and parameters of their choosing. This constitutes a cross-site request forgery (CSRF) vulnerability with high integrity impact on the MISP instance\u0027s security-related data.\u0026nbsp;\u003c/p\u003e\u003cp\u003eThe vulnerability was identified during an internal security review and was not externally reported. The fix is included in MISP v2.5.47.\u003c/p\u003e"
}
],
"value": "MISP\u0027s WorkflowsController exposed the moduleStatelessExecution action in the Security component\u0027s unlockedActions list. In CakePHP, listing an action in unlockedActions disables both the CSRF token check and the field hash validation for that action. Because moduleStatelessExecution executes a workflow module\u0027s exec() method with caller-supplied input and parameters, the absence of CSRF protection allowed an attacker to craft a cross-site form post (or equivalent cross-origin request) that, when submitted by an authenticated site administrator, would cause the administrator\u0027s browser to invoke the action on the MISP instance.\u00a0\n\nThe attacker could select any workflow module to execute, including action modules that write blocklist and warninglist entries, and supply arbitrary input and parameters of their choosing. This constitutes a cross-site request forgery (CSRF) vulnerability with high integrity impact on the MISP instance\u0027s security-related data.\u00a0\n\nThe vulnerability was identified during an internal security review and was not externally reported. The fix is included in MISP v2.5.47."
}
],
"impacts": [
{
"capecId": "CAPEC-124",
"descriptions": [
{
"lang": "en",
"value": "CAPEC-124 Cross-Site Request Forgery"
}
]
}
],
"metrics": [
{
"cvssV4_0": {
"Automatable": "NOT_DEFINED",
"Recovery": "NOT_DEFINED",
"Safety": "NOT_DEFINED",
"attackComplexity": "LOW",
"attackRequirements": "NONE",
"attackVector": "NETWORK",
"baseScore": 6.9,
"baseSeverity": "MEDIUM",
"privilegesRequired": "NONE",
"providerUrgency": "NOT_DEFINED",
"subAvailabilityImpact": "NONE",
"subConfidentialityImpact": "NONE",
"subIntegrityImpact": "NONE",
"userInteraction": "ACTIVE",
"valueDensity": "NOT_DEFINED",
"vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:A/VC:N/VI:H/VA:N/SC:N/SI:N/SA:N",
"version": "4.0",
"vulnAvailabilityImpact": "NONE",
"vulnConfidentialityImpact": "NONE",
"vulnIntegrityImpact": "HIGH",
"vulnerabilityResponseEffort": "NOT_DEFINED"
},
"format": "CVSS",
"scenarios": [
{
"lang": "en",
"value": "GENERAL"
}
]
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-352",
"description": "CWE-352 Cross-Site Request Forgery (CSRF)",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"orgId": "00000000-0000-4000-9000-000000000000"
},
"references": [
{
"name": "Security patch",
"tags": [
"patch"
],
"url": "https://github.com/MISP/MISP/commit/f05e0d1c2"
}
],
"solutions": [
{
"lang": "en",
"supportingMedia": [
{
"base64": false,
"type": "text/html",
"value": "\u003cp\u003eThe moduleStatelessExecution action is removed from the unlockedActions list, restoring the CSRF token check. A new mechanism (_csrfTokenHeaderOnly) is introduced that enforces the CSRF token while exempting only the field hash, which the legitimate caller (the module dialog) cannot produce because it posts a hand-built object. The module dialog\u0027s AJAX request is updated to send the CSRF token in the X-CSRF-Token header, satisfying the token check without requiring a field hash.\u003c/p\u003e"
}
],
"value": "The moduleStatelessExecution action is removed from the unlockedActions list, restoring the CSRF token check. A new mechanism (_csrfTokenHeaderOnly) is introduced that enforces the CSRF token while exempting only the field hash, which the legitimate caller (the module dialog) cannot produce because it posts a hand-built object. The module dialog\u0027s AJAX request is updated to send the CSRF token in the X-CSRF-Token header, satisfying the token check without requiring a field hash."
}
],
"title": "MISP CSRF vulnerability in workflow moduleStatelessExecution allows cross-site execution of workflow modules",
"x_gcve": [
{
"extensions": {
"bcp-05-x-01": {
"ai_annotations": [
{
"ai_level": "generated",
"description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
"gna_source": 1,
"models": [
{
"gna_source": 1,
"identifier": "qwen3.8:27b",
"name": "qwen3.8:27b",
"source": "ollama"
}
],
"review_status": "full",
"scope": "record",
"tags": [
"ai-computer-assisted:llm-generated",
"ai-computer-assisted:classification"
]
}
]
},
"bcp-05-x-02": {
"x_patch2vuln": {
"assumptions": [
"The affected version range is inferred from the tag_version_boundary metadata (v2.5.47, 6 commits after fix), implying the fix is included in v2.5.47 and prior versions are affected. The exact introduction version of the vulnerable code is unknown.",
"UI:A is used because the CSRF attack requires the victim to actively navigate to the attacker\u0027s page or click a link; a purely passive load (e.g., embedded iframe auto-submit) would map to UI:P, but the commit message describes a \u0027cross-site form post\u0027 which typically requires at least a page visit.",
"SI:N is chosen because the evidence only demonstrates integrity impact on MISP\u0027s own blocklist and warninglist data; no evidence supports impact on downstream or external systems.",
"The CAPEC-124 mapping is confident based on the explicit description of a cross-site form post in the commit message.",
"The commit message states the issue was \u0027found during the internal review, not externally reported,\u0027 but does not explicitly name a finder; therefore no finder credit is assigned.",
"The _csrfTokenHeaderOnly method is referenced in the patch but its implementation is not shown; it is assumed to enforce CSRF token validation while exempting the field hash, as described in the commit message."
],
"capecRationale": [
{
"capecId": "CAPEC-124",
"rationale": "The attack pattern described in the commit message is a cross-site form post that causes an authenticated site admin to execute a workflow module with attacker-chosen parameters. This maps directly to CAPEC-124, which describes an attacker exploiting the trust a web application places in a user\u0027s browser session to perform unauthorized actions. The mapping is confident because the commit message explicitly describes a cross-site form post bypassing CSRF protection."
}
],
"commit": "f05e0d1c263b211108644fe6f1600ad7fde9b915",
"confidence": "high",
"credits": [
{
"lang": "en",
"type": "remediation developer",
"value": "iglocska"
},
{
"lang": "en",
"type": "remediation developer",
"value": "Claude Opus 5 (1M context)"
}
],
"cvssRationale": "AV:N: The attack is delivered over the network via a malicious web page. AC:L: Crafting a cross-site form post is straightforward. AT:N: No special conditions or timing are required. PR:N: The attacker requires no account or privileges on the MISP instance. UI:A: The victim (a site admin) must actively visit the attacker-controlled page or click a link that triggers the form submission. VC:N: No confidentiality impact is evident. VI:H: The attacker can write arbitrary blocklist and warninglist entries, significantly compromising the integrity of MISP\u0027s security data. VA:N: No availability impact. SC/SI/SA:N: No evidence of impact on other systems beyond MISP itself.",
"fixSummary": "The moduleStatelessExecution action is removed from the unlockedActions list, restoring the CSRF token check. A new mechanism (_csrfTokenHeaderOnly) is introduced that enforces the CSRF token while exempting only the field hash, which the legitimate caller (the module dialog) cannot produce because it posts a hand-built object. The module dialog\u0027s AJAX request is updated to send the CSRF token in the X-CSRF-Token header, satisfying the token check without requiring a field hash.",
"generatedAt": "2026-09-22T12:15:58.553959Z",
"generator": "patch2vuln.py",
"model": "qwen3.8:27b",
"modelComparison": {
"rankings": [
{
"agreementScore": 9,
"assumptionCount": 6,
"confidence": "high",
"model": "qwen3.8:27b",
"score": 5
}
],
"selectedModel": "qwen3.8:27b",
"selectionMethod": "deterministic-consensus-v1",
"selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
},
"patchSha256": "3f8a00ca29761ed6a3986633e60ec67a964e3bfc1dc1e2631a2c2d80df827d30",
"patchSummary": "In WorkflowsController::beforeFilter(), the line adding \u0027moduleStatelessExecution\u0027 to $this-\u003eSecurity-\u003eunlockedActions is replaced with a call to $this-\u003e_csrfTokenHeaderOnly([\u0027moduleStatelessExecution\u0027]), which enforces CSRF token validation while waiving only the field hash requirement. In execute_module.ctp, the jQuery $.ajax() call is modified to include the header \u0027X-CSRF-Token\u0027 populated from window.csrfToken, so the legitimate client-side caller passes the token check via the header rather than a form field.",
"patchTruncated": false,
"patches": [
{
"commit": "f05e0d1c263b211108644fe6f1600ad7fde9b915",
"patchSha256": "3f8a00ca29761ed6a3986633e60ec67a964e3bfc1dc1e2631a2c2d80df827d30",
"source": "https://github.com/MISP/MISP/commit/f05e0d1c2.patch",
"sourceUrl": "https://github.com/MISP/MISP/commit/f05e0d1c2.patch",
"subject": "fix: [security] Keep the CSRF check on stateless workflow"
}
],
"source": "https://github.com/MISP/MISP/commit/f05e0d1c2.patch",
"subject": "fix: [security] Keep the CSRF check on stateless workflow",
"tagVersionBoundary": {
"commits_after_fix": 6,
"repository": "https://github.com/MISP/MISP",
"tag": "v2.5.47",
"version": "2.5.47",
"version_type": "semver"
},
"weaknessRationale": [
{
"cweId": "CWE-352",
"rationale": "The vulnerable action was placed in unlockedActions, which disabled the CSRF token check. This allowed a cross-site form post to trigger a state-changing action (workflow module execution) on behalf of an authenticated administrator, which is the textbook definition of CSRF."
}
]
}
}
},
"recordType": "advisory",
"vulnId": "GCVE-1-2026-20261"
}
]
}
},
"cveMetadata": {
"assignerOrgId": "00000000-0000-4000-9000-000000000000",
"cveId": "CVE-2026-95658",
"datePublished": "2026-09-22T12:31:11.850218Z",
"dateReserved": "2026-09-22T12:31:20.344Z",
"dateUpdated": "2026-09-22T12:31:20.402691Z",
"state": "PUBLISHED",
"vulnId": "GCVE-1-2026-20261"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.1"
}
GCVE-1-2026-20153 (CVE-2026-93295)
Vulnerability from gna-1 – Published: 2026-09-17 16:23 – Updated: 2026-09-22 07:42| URL | Tags |
|---|---|
| https://github.com/MISP/misp/commit/120813344 | patch |
qwen3.8:27b
advisory
bcp-05-x-01bcp-05-x-02
Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.
| Model | Source | Identifier |
|---|---|---|
| qwen3.8:27b | ollama | qwen3.8:27b |
- Generator
-
patch2vuln.pyon 2026-09-17 16:22 - Model
qwen3.8:27b- Input
-
https://github.com/MISP/misp/commit/120813344.patch
8e176aa69e4b… - Confidence
- high
| Commit | Subject | Patch SHA-256 |
|---|---|---|
120813344f36
|
fix: [security] Refuse console path switches in background | 8e176aa69e4b… |
Fix summary
The fix introduces a validation layer in BackgroundJobsTool::enqueue() that rejects any job argument matching a reserved CakePHP console path switch (-app, --app, -working, --working, -root, --root, -webroot, --webroot), throwing an InvalidArgumentException before the job is queued. Additionally, the EventsController::contact() method now casts the person field to a boolean, preventing it from ever forming a switch/value pair with the adjacent message field. Together, these changes ensure that user-supplied data can never be interpreted as a console path directive by the worker process.
Patch summary
Three files are modified. In BackgroundJobsTool.php, a new RESERVED_ARGUMENTS constant lists the eight console path switches, and a new private validateArgs() method iterates over job arguments and throws InvalidArgumentException if any string argument exactly matches a reserved switch. The enqueue() method calls validateArgs() before dispatching. In EventsController.php, the person field assignment in contact() is changed from a raw value to (bool) cast. A new test file BackgroundJobsToolArgumentsTest.php verifies that ordinary arguments pass, free text merely containing a switch substring passes, every reserved switch is refused, and non-string arguments are unaffected.
CVSS rationale
AV:N: The contact endpoint is reachable over the network via HTTP. AC:L: The attack requires only crafting two form fields with a known switch string and a phar:// path; no race conditions or special timing. AT:N: No specific target configuration or state is required beyond a standard MISP installation. PR:L: The attacker needs a low-privilege authenticated MISP account with access to the event contact action. UI:N: No victim interaction is needed beyond the attacker's own form submission. VC:H, VI:H, VA:H: Successful exploitation results in arbitrary PHP code execution as the web user, giving full control over the MISP server's data, processes, and availability. SC:N, SI:N, SA:N: The impact is confined to the MISP server itself; no evidence of lateral impact on other systems.
Weakness rationale
- CWE-74 User-controlled form fields (person, message) are passed as arguments to a downstream component (the CakePHP console dispatcher) where special elements (path switches like -app, -root) are interpreted as directives rather than data. The lack of neutralization of these special elements in the job argument vector is the root cause.
- CWE-20 The background job enqueue path accepted arbitrary string arguments without validating them against the set of values that the downstream console dispatcher interprets as control switches. The fix adds explicit validation to reject reserved argument values.
Attack pattern rationale
- CAPEC-14 The attacker manipulates the person and message parameters of the events/contact form to inject CakePHP console path switches into the background job argv. The parameters are modified so that the downstream console dispatcher interprets them as directives (setting the application root) rather than as data. This is the closest CAPEC pattern; the attack does not involve classic OS command injection (CAPEC-88) but rather argument injection into a framework-level dispatcher, making CAPEC-14 the best available match.
Assumptions to verify
- The affected version boundary is inferred from the tag_version_boundary metadata (v2.5.47, 43 commits after fix); the exact first fixed release is not explicitly stated in the patch and may differ.
- PR:L assumes the events/contact endpoint requires at least a low-privilege authenticated session; if the endpoint is accessible without authentication in some MISP configurations, PR should be U (Unprivileged).
- The CAPEC-14 mapping is the closest available pattern; the attack is more specifically an argument injection into a framework dispatcher rather than generic parameter manipulation, but no CAPEC entry precisely describes this mechanism.
- The phar:// exploitation path assumes the target PHP process has the phar stream wrapper enabled and that the attacker can place or reference a crafted archive accessible via the phar:// URI; the patch does not explicitly confirm the phar wrapper is enabled by default.
- The commit date (17 Sep 2026) is in the future relative to typical CVE timelines; this is taken at face value from the patch metadata.
Model comparison
Selected qwen3.8:27b
by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.
| Model | Score | Agreement | Confidence | Assumptions |
|---|---|---|---|---|
qwen3.8:27b |
6 | 9 | high | 5 |
{
"containers": {
"cna": {
"affected": [
{
"defaultStatus": "unaffected",
"modules": [
"app/Controller/EventsController.php",
"app/Lib/Tools/BackgroundJobsTool.php"
],
"product": "misp",
"repo": "https://github.com/MISP/misp",
"vendor": "misp",
"versions": [
{
"lessThan": "2.5.47",
"status": "affected",
"version": "\u003c 2.5.47",
"versionType": "semver"
}
]
}
],
"credits": [
{
"lang": "en",
"type": "reporter",
"value": "Niels Teusink of Eye Security"
},
{
"lang": "en",
"type": "remediation developer",
"value": "iglocska"
},
{
"lang": "en",
"type": "remediation developer",
"value": "Claude Opus 5 (1M context)"
}
],
"descriptions": [
{
"lang": "en",
"supportingMedia": [
{
"base64": false,
"type": "text/html",
"value": "\u003cp\u003eMISP contains a vulnerability in its background job dispatch mechanism that allows remote code execution as the web user. Background job arguments are passed directly as the argv of the CakePHP console process. CakePHP\u0027s ShellDispatcher::_parsePaths() scans the entire argv for path switches (-app, --app, -working, --working, -root, --root, -webroot, --webroot) and uses the following element as the application root. The events/contact endpoint passes user-controlled fields (message and person) into job arguments without validation. An attacker who can submit the contact form can set the person field to a reserved switch and the message field to a phar:// URI pointing to a malicious archive. The CakePHP bootstrap then includes Config/core.php from within that archive, executing attacker-controlled PHP code with the privileges of the web user.\u0026nbsp;\u003c/p\u003e\u003cp\u003eThe vulnerability requires the ability to submit the events/contact form (or any other endpoint that forwards user input into background job arguments). No special timing or race condition is required; the attack is deterministic once the crafted parameters are accepted. The impact is full remote code execution in the context of the MISP web server process, allowing data exfiltration, persistence, and lateral movement within the host.\u003c/p\u003e"
}
],
"value": "MISP contains a vulnerability in its background job dispatch mechanism that allows remote code execution as the web user. Background job arguments are passed directly as the argv of the CakePHP console process. CakePHP\u0027s ShellDispatcher::_parsePaths() scans the entire argv for path switches (-app, --app, -working, --working, -root, --root, -webroot, --webroot) and uses the following element as the application root. The events/contact endpoint passes user-controlled fields (message and person) into job arguments without validation. An attacker who can submit the contact form can set the person field to a reserved switch and the message field to a phar:// URI pointing to a malicious archive. The CakePHP bootstrap then includes Config/core.php from within that archive, executing attacker-controlled PHP code with the privileges of the web user.\u00a0\n\nThe vulnerability requires the ability to submit the events/contact form (or any other endpoint that forwards user input into background job arguments). No special timing or race condition is required; the attack is deterministic once the crafted parameters are accepted. The impact is full remote code execution in the context of the MISP web server process, allowing data exfiltration, persistence, and lateral movement within the host."
}
],
"impacts": [
{
"capecId": "CAPEC-14",
"descriptions": [
{
"lang": "en",
"value": "CAPEC-14 Parameter Manipulation"
}
]
}
],
"metrics": [
{
"cvssV4_0": {
"Automatable": "NOT_DEFINED",
"Recovery": "NOT_DEFINED",
"Safety": "NOT_DEFINED",
"attackComplexity": "LOW",
"attackRequirements": "NONE",
"attackVector": "NETWORK",
"baseScore": 8.7,
"baseSeverity": "HIGH",
"privilegesRequired": "LOW",
"providerUrgency": "NOT_DEFINED",
"subAvailabilityImpact": "NONE",
"subConfidentialityImpact": "NONE",
"subIntegrityImpact": "NONE",
"userInteraction": "NONE",
"valueDensity": "NOT_DEFINED",
"vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N",
"version": "4.0",
"vulnAvailabilityImpact": "HIGH",
"vulnConfidentialityImpact": "HIGH",
"vulnIntegrityImpact": "HIGH",
"vulnerabilityResponseEffort": "NOT_DEFINED"
},
"format": "CVSS",
"scenarios": [
{
"lang": "en",
"value": "GENERAL"
}
]
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-74",
"description": "CWE-74 Improper Neutralization of Special Elements in Output Used by a Downstream Component (\u0027Injection\u0027)",
"lang": "en",
"type": "CWE"
}
]
},
{
"descriptions": [
{
"cweId": "CWE-20",
"description": "CWE-20 Improper Input Validation",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"orgId": "00000000-0000-4000-9000-000000000000"
},
"references": [
{
"name": "Security patch",
"tags": [
"patch"
],
"url": "https://github.com/MISP/misp/commit/120813344"
}
],
"solutions": [
{
"lang": "en",
"supportingMedia": [
{
"base64": false,
"type": "text/html",
"value": "\u003cp\u003eThe fix introduces a validation layer in BackgroundJobsTool::enqueue() that rejects any job argument matching a reserved CakePHP console path switch (-app, --app, -working, --working, -root, --root, -webroot, --webroot), throwing an InvalidArgumentException before the job is queued. Additionally, the EventsController::contact() method now casts the person field to a boolean, preventing it from ever forming a switch/value pair with the adjacent message field. Together, these changes ensure that user-supplied data can never be interpreted as a console path directive by the worker process.\u003c/p\u003e"
}
],
"value": "The fix introduces a validation layer in BackgroundJobsTool::enqueue() that rejects any job argument matching a reserved CakePHP console path switch (-app, --app, -working, --working, -root, --root, -webroot, --webroot), throwing an InvalidArgumentException before the job is queued. Additionally, the EventsController::contact() method now casts the person field to a boolean, preventing it from ever forming a switch/value pair with the adjacent message field. Together, these changes ensure that user-supplied data can never be interpreted as a console path directive by the worker process."
}
],
"source": {
"discovery": "UNKNOWN"
},
"title": "MISP Background Job Argument Injection via Console Path Switches Enables Remote Code Execution",
"x_gcve": [
{
"extensions": {
"bcp-05-x-01": {
"ai_annotations": [
{
"ai_level": "generated",
"description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
"gna_source": 1,
"models": [
{
"gna_source": 1,
"identifier": "qwen3.8:27b",
"name": "qwen3.8:27b",
"source": "ollama"
}
],
"review_status": "full",
"scope": "record",
"tags": [
"ai-computer-assisted:llm-generated",
"ai-computer-assisted:classification"
]
}
]
},
"bcp-05-x-02": {
"x_patch2vuln": {
"assumptions": [
"The affected version boundary is inferred from the tag_version_boundary metadata (v2.5.47, 43 commits after fix); the exact first fixed release is not explicitly stated in the patch and may differ.",
"PR:L assumes the events/contact endpoint requires at least a low-privilege authenticated session; if the endpoint is accessible without authentication in some MISP configurations, PR should be U (Unprivileged).",
"The CAPEC-14 mapping is the closest available pattern; the attack is more specifically an argument injection into a framework dispatcher rather than generic parameter manipulation, but no CAPEC entry precisely describes this mechanism.",
"The phar:// exploitation path assumes the target PHP process has the phar stream wrapper enabled and that the attacker can place or reference a crafted archive accessible via the phar:// URI; the patch does not explicitly confirm the phar wrapper is enabled by default.",
"The commit date (17 Sep 2026) is in the future relative to typical CVE timelines; this is taken at face value from the patch metadata."
],
"capecRationale": [
{
"capecId": "CAPEC-14",
"rationale": "The attacker manipulates the person and message parameters of the events/contact form to inject CakePHP console path switches into the background job argv. The parameters are modified so that the downstream console dispatcher interprets them as directives (setting the application root) rather than as data. This is the closest CAPEC pattern; the attack does not involve classic OS command injection (CAPEC-88) but rather argument injection into a framework-level dispatcher, making CAPEC-14 the best available match."
}
],
"commit": "120813344f36b8fad15cd917e97b09fa3497156b",
"confidence": "high",
"credits": [
{
"lang": "en",
"type": "reporter",
"value": "Niels Teusink of Eye Security"
},
{
"lang": "en",
"type": "remediation developer",
"value": "iglocska"
},
{
"lang": "en",
"type": "remediation developer",
"value": "Claude Opus 5 (1M context)"
}
],
"cvssRationale": "AV:N: The contact endpoint is reachable over the network via HTTP. AC:L: The attack requires only crafting two form fields with a known switch string and a phar:// path; no race conditions or special timing. AT:N: No specific target configuration or state is required beyond a standard MISP installation. PR:L: The attacker needs a low-privilege authenticated MISP account with access to the event contact action. UI:N: No victim interaction is needed beyond the attacker\u0027s own form submission. VC:H, VI:H, VA:H: Successful exploitation results in arbitrary PHP code execution as the web user, giving full control over the MISP server\u0027s data, processes, and availability. SC:N, SI:N, SA:N: The impact is confined to the MISP server itself; no evidence of lateral impact on other systems.",
"fixSummary": "The fix introduces a validation layer in BackgroundJobsTool::enqueue() that rejects any job argument matching a reserved CakePHP console path switch (-app, --app, -working, --working, -root, --root, -webroot, --webroot), throwing an InvalidArgumentException before the job is queued. Additionally, the EventsController::contact() method now casts the person field to a boolean, preventing it from ever forming a switch/value pair with the adjacent message field. Together, these changes ensure that user-supplied data can never be interpreted as a console path directive by the worker process.",
"generatedAt": "2026-09-17T16:22:44.930762Z",
"generator": "patch2vuln.py",
"model": "qwen3.8:27b",
"modelComparison": {
"rankings": [
{
"agreementScore": 9,
"assumptionCount": 5,
"confidence": "high",
"model": "qwen3.8:27b",
"score": 6
}
],
"selectedModel": "qwen3.8:27b",
"selectionMethod": "deterministic-consensus-v1",
"selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
},
"patchSha256": "8e176aa69e4b6c13f3909225292a96c127076f333a6a49c1824599a79533de77",
"patchSummary": "Three files are modified. In BackgroundJobsTool.php, a new RESERVED_ARGUMENTS constant lists the eight console path switches, and a new private validateArgs() method iterates over job arguments and throws InvalidArgumentException if any string argument exactly matches a reserved switch. The enqueue() method calls validateArgs() before dispatching. In EventsController.php, the person field assignment in contact() is changed from a raw value to (bool) cast. A new test file BackgroundJobsToolArgumentsTest.php verifies that ordinary arguments pass, free text merely containing a switch substring passes, every reserved switch is refused, and non-string arguments are unaffected.",
"patchTruncated": false,
"patches": [
{
"commit": "120813344f36b8fad15cd917e97b09fa3497156b",
"patchSha256": "8e176aa69e4b6c13f3909225292a96c127076f333a6a49c1824599a79533de77",
"source": "https://github.com/MISP/misp/commit/120813344.patch",
"sourceUrl": "https://github.com/MISP/misp/commit/120813344.patch",
"subject": "fix: [security] Refuse console path switches in background"
}
],
"source": "https://github.com/MISP/misp/commit/120813344.patch",
"subject": "fix: [security] Refuse console path switches in background",
"tagVersionBoundary": {
"commits_after_fix": 43,
"repository": "https://github.com/MISP/misp",
"tag": "v2.5.47",
"version": "2.5.47",
"version_type": "semver"
},
"weaknessRationale": [
{
"cweId": "CWE-74",
"rationale": "User-controlled form fields (person, message) are passed as arguments to a downstream component (the CakePHP console dispatcher) where special elements (path switches like -app, -root) are interpreted as directives rather than data. The lack of neutralization of these special elements in the job argument vector is the root cause."
},
{
"cweId": "CWE-20",
"rationale": "The background job enqueue path accepted arbitrary string arguments without validating them against the set of values that the downstream console dispatcher interprets as control switches. The fix adds explicit validation to reject reserved argument values."
}
]
}
}
},
"recordType": "advisory",
"vulnId": "gcve-1-2026-20153"
}
],
"x_generator": {
"engine": "Vulnogram 0.2.0"
}
}
},
"cveMetadata": {
"assignerOrgId": "00000000-0000-4000-9000-000000000000",
"cveId": "CVE-2026-93295",
"datePublished": "2026-09-17T16:23:00.000Z",
"dateReserved": "2026-09-17T16:25:00.000Z",
"dateUpdated": "2026-09-22T07:42:03.681655Z",
"requesterUserId": "00000000-0000-4000-9000-000000000000",
"serial": 1,
"state": "PUBLISHED",
"vulnId": "gcve-1-2026-20153"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.1"
}
GCVE-1-2026-20206 (CVE-2026-93296)
Vulnerability from gna-1 – Published: 2026-09-17 16:29 – Updated: 2026-09-22 07:36- CWE-79 - Improper Neutralization of Input During Web Page Generation ('Cross-site Scripting')
| URL | Tags |
|---|---|
| https://github.com/MISP/misp/commit/382188d2f | patch |
qwen3.8:27b
advisory
bcp-05-x-01bcp-05-x-02
Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.
| Model | Source | Identifier |
|---|---|---|
| qwen3.8:27b | ollama | qwen3.8:27b |
- Generator
-
patch2vuln.pyon 2026-09-17 16:28 - Model
qwen3.8:27b- Input
-
https://github.com/MISP/misp/commit/382188d2f.patch
3c6f0376c712… - Confidence
- high
| Commit | Subject | Patch SHA-256 |
|---|---|---|
382188d2f270
|
fix: [security] Escape object names in the Overmind | 3c6f0376c712… |
Fix summary
The fix applies HTML output encoding to the user-controlled label variable (lbl) before it is interpolated into the innerHTML string in both affected view templates. By wrapping the label with escapeHtml(), any HTML metacharacters in object names are neutralized, preventing injected markup from being interpreted as executable HTML or JavaScript in the viewer's browser.
Patch summary
Two one-line changes in the Overmind theme view templates. In app/View/Themed/Overmind/Elements/Events/View/event_general.ctp (line 607) and app/View/Themed/Overmind/Elements/Servers/View/preview_general.ctp (line 323), the raw lbl variable used in the title attribute and inner text of the legend span element is replaced with escapeHtml(lbl). No other logic, routing, or configuration changes are made.
CVSS rationale
AV:N: The vulnerability is exploitable over the network via a web browser. AC:L: Exploitation requires only creating an object with a crafted name; no race or special conditions. AT:N: The attack targets the same MISP instance. PR:L: The attacker needs low-level authenticated access to create or rename an MISP object. UI:P: The victim passively interacts by viewing the Overmind dashboard page; no click or explicit action is needed beyond normal page rendering. VC/VI/VA:N: No direct impact on the MISP server's confidentiality, integrity, or availability. SC:H/SI:H: The injected script executes in the victim's browser within the MISP origin, enabling full session compromise (cookie theft, data exfiltration, arbitrary actions). SA:N: No availability impact on the victim's browser.
Weakness rationale
- CWE-79 User-controlled object names are concatenated directly into an innerHTML string without HTML encoding, allowing injected markup to execute in the victim's browser. This is a textbook stored XSS via missing output encoding at an HTML sink.
Attack pattern rationale
- CAPEC-1 The attack pattern involves injecting client-side script into a web application via user-controllable data (object names) that is rendered without encoding in an HTML context. CAPEC-1 is the canonical XSS attack pattern and directly matches the mechanism: attacker-controlled data flows into an innerHTML sink. No uncertainty in this mapping; the patch explicitly describes the innerHTML concatenation of unescaped labels.
Assumptions to verify
- The exact affected version range is not explicitly stated in the patch; the tag_version_boundary indicates the fix commit is 55 commits after v2.5.47, suggesting versions up to (but not including) the release containing this commit are affected. The precise fixed version is unspecified.
- PR:L assumes the attacker needs at least low-level authenticated access to create or rename a MISP object whose name appears in the Overmind legend. If unauthenticated object creation is possible in some deployment, PR could be N.
- UI:P assumes the victim simply views the Overmind dashboard page (passive interaction). If the page requires a specific navigation action beyond normal browsing, UI could be A.
- SC:H and SI:H assume the XSS grants full access to the victim's MISP session (cookies, API tokens, in-page data). A more conservative assessment could rate these L if the Overmind view exposes limited data.
- The CAPEC-1 mapping is direct and unambiguous for this stored XSS; no uncertainty is noted beyond the standard CAPEC granularity.
Model comparison
Selected qwen3.8:27b
by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.
| Model | Score | Agreement | Confidence | Assumptions |
|---|---|---|---|---|
qwen3.8:27b |
6 | 9 | high | 5 |
{
"containers": {
"cna": {
"affected": [
{
"defaultStatus": "unaffected",
"modules": [
"Overmind theme - Events View (event_general.ctp)",
"Overmind theme - Servers View (preview_general.ctp)"
],
"product": "misp",
"programFiles": [
"app/View/Themed/Overmind/Elements/Events/View/event_general.ctp",
"app/View/Themed/Overmind/Elements/Servers/View/preview_general.ctp"
],
"repo": "https://github.com/MISP/misp",
"vendor": "misp",
"versions": [
{
"lessThan": "2.5.47",
"status": "affected",
"version": "unspecified",
"versionType": "semver"
}
]
}
],
"credits": [
{
"lang": "en",
"type": "reporter",
"value": "Jeroen Pinoy"
},
{
"lang": "en",
"type": "remediation developer",
"value": "iglocska"
},
{
"lang": "en",
"type": "remediation developer",
"value": "Claude Opus 4.8"
}
],
"descriptions": [
{
"lang": "en",
"supportingMedia": [
{
"base64": false,
"type": "text/html",
"value": "\u003cp\u003eMISP contains a stored cross-site scripting (XSS) vulnerability in the Overmind theme\u0027s statistics views. The event General card and the server/feed preview card constructed donut chart legend labels by directly concatenating object name or category keys into an innerHTML string without HTML-encoding. Because MISP object names are user-controllable by any authenticated user with sufficient permissions to create or modify such objects, an attacker could craft a name containing malicious markup. When any other user viewed the affected Overmind dashboard, the injected markup would be interpreted as live HTML/JavaScript in the victim\u0027s browser, executing in the context of the MISP application origin. This could allow session hijacking, data exfiltration, or arbitrary actions performed on behalf of the victim.\u0026nbsp;\u003c/p\u003e\u003cp\u003eThe vulnerability requires the attacker to have low-level authenticated access to create or rename an object whose name is rendered in the legend, and the victim to view the Overmind event or server preview page. No special browser conditions or race conditions are required.\u003c/p\u003e"
}
],
"value": "MISP contains a stored cross-site scripting (XSS) vulnerability in the Overmind theme\u0027s statistics views. The event General card and the server/feed preview card constructed donut chart legend labels by directly concatenating object name or category keys into an innerHTML string without HTML-encoding. Because MISP object names are user-controllable by any authenticated user with sufficient permissions to create or modify such objects, an attacker could craft a name containing malicious markup. When any other user viewed the affected Overmind dashboard, the injected markup would be interpreted as live HTML/JavaScript in the victim\u0027s browser, executing in the context of the MISP application origin. This could allow session hijacking, data exfiltration, or arbitrary actions performed on behalf of the victim.\u00a0\n\nThe vulnerability requires the attacker to have low-level authenticated access to create or rename an object whose name is rendered in the legend, and the victim to view the Overmind event or server preview page. No special browser conditions or race conditions are required."
}
],
"impacts": [
{
"capecId": "CAPEC-1",
"descriptions": [
{
"lang": "en",
"value": "CAPEC-1 Cross Site Scripting"
}
]
}
],
"metrics": [
{
"cvssV4_0": {
"Automatable": "NOT_DEFINED",
"Recovery": "NOT_DEFINED",
"Safety": "NOT_DEFINED",
"attackComplexity": "LOW",
"attackRequirements": "NONE",
"attackVector": "NETWORK",
"baseScore": 8.5,
"baseSeverity": "HIGH",
"privilegesRequired": "LOW",
"providerUrgency": "NOT_DEFINED",
"subAvailabilityImpact": "NONE",
"subConfidentialityImpact": "LOW",
"subIntegrityImpact": "LOW",
"userInteraction": "PASSIVE",
"valueDensity": "NOT_DEFINED",
"vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:P/VC:H/VI:H/VA:N/SC:L/SI:L/SA:N",
"version": "4.0",
"vulnAvailabilityImpact": "NONE",
"vulnConfidentialityImpact": "HIGH",
"vulnIntegrityImpact": "HIGH",
"vulnerabilityResponseEffort": "NOT_DEFINED"
},
"format": "CVSS",
"scenarios": [
{
"lang": "en",
"value": "GENERAL"
}
]
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-79",
"description": "CWE-79 Improper Neutralization of Input During Web Page Generation (\u0027Cross-site Scripting\u0027)",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"orgId": "00000000-0000-4000-9000-000000000000"
},
"references": [
{
"name": "Security patch",
"tags": [
"patch"
],
"url": "https://github.com/MISP/misp/commit/382188d2f"
}
],
"solutions": [
{
"lang": "en",
"supportingMedia": [
{
"base64": false,
"type": "text/html",
"value": "\u003cp\u003eThe fix applies HTML output encoding to the user-controlled label variable (lbl) before it is interpolated into the innerHTML string in both affected view templates. By wrapping the label with escapeHtml(), any HTML metacharacters in object names are neutralized, preventing injected markup from being interpreted as executable HTML or JavaScript in the viewer\u0027s browser.\u003c/p\u003e"
}
],
"value": "The fix applies HTML output encoding to the user-controlled label variable (lbl) before it is interpolated into the innerHTML string in both affected view templates. By wrapping the label with escapeHtml(), any HTML metacharacters in object names are neutralized, preventing injected markup from being interpreted as executable HTML or JavaScript in the viewer\u0027s browser."
}
],
"source": {
"discovery": "UNKNOWN"
},
"title": "MISP Overmind: Stored Cross-Site Scripting via Unescaped Object Names in Statistics Legends",
"x_gcve": [
{
"extensions": {
"bcp-05-x-01": {
"ai_annotations": [
{
"ai_level": "generated",
"description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
"gna_source": 1,
"models": [
{
"gna_source": 1,
"identifier": "qwen3.8:27b",
"name": "qwen3.8:27b",
"source": "ollama"
}
],
"review_status": "full",
"scope": "record",
"tags": [
"ai-computer-assisted:llm-generated",
"ai-computer-assisted:classification"
]
}
]
},
"bcp-05-x-02": {
"x_patch2vuln": {
"assumptions": [
"The exact affected version range is not explicitly stated in the patch; the tag_version_boundary indicates the fix commit is 55 commits after v2.5.47, suggesting versions up to (but not including) the release containing this commit are affected. The precise fixed version is unspecified.",
"PR:L assumes the attacker needs at least low-level authenticated access to create or rename a MISP object whose name appears in the Overmind legend. If unauthenticated object creation is possible in some deployment, PR could be N.",
"UI:P assumes the victim simply views the Overmind dashboard page (passive interaction). If the page requires a specific navigation action beyond normal browsing, UI could be A.",
"SC:H and SI:H assume the XSS grants full access to the victim\u0027s MISP session (cookies, API tokens, in-page data). A more conservative assessment could rate these L if the Overmind view exposes limited data.",
"The CAPEC-1 mapping is direct and unambiguous for this stored XSS; no uncertainty is noted beyond the standard CAPEC granularity."
],
"capecRationale": [
{
"capecId": "CAPEC-1",
"rationale": "The attack pattern involves injecting client-side script into a web application via user-controllable data (object names) that is rendered without encoding in an HTML context. CAPEC-1 is the canonical XSS attack pattern and directly matches the mechanism: attacker-controlled data flows into an innerHTML sink. No uncertainty in this mapping; the patch explicitly describes the innerHTML concatenation of unescaped labels."
}
],
"commit": "382188d2f27004118592cf32f970313b31abcb52",
"confidence": "high",
"credits": [
{
"lang": "en",
"type": "reporter",
"value": "Jeroen Pinoy"
},
{
"lang": "en",
"type": "remediation developer",
"value": "iglocska"
},
{
"lang": "en",
"type": "remediation developer",
"value": "Claude Opus 4.8"
}
],
"cvssRationale": "AV:N: The vulnerability is exploitable over the network via a web browser. AC:L: Exploitation requires only creating an object with a crafted name; no race or special conditions. AT:N: The attack targets the same MISP instance. PR:L: The attacker needs low-level authenticated access to create or rename an MISP object. UI:P: The victim passively interacts by viewing the Overmind dashboard page; no click or explicit action is needed beyond normal page rendering. VC/VI/VA:N: No direct impact on the MISP server\u0027s confidentiality, integrity, or availability. SC:H/SI:H: The injected script executes in the victim\u0027s browser within the MISP origin, enabling full session compromise (cookie theft, data exfiltration, arbitrary actions). SA:N: No availability impact on the victim\u0027s browser.",
"fixSummary": "The fix applies HTML output encoding to the user-controlled label variable (lbl) before it is interpolated into the innerHTML string in both affected view templates. By wrapping the label with escapeHtml(), any HTML metacharacters in object names are neutralized, preventing injected markup from being interpreted as executable HTML or JavaScript in the viewer\u0027s browser.",
"generatedAt": "2026-09-17T16:28:27.480704Z",
"generator": "patch2vuln.py",
"model": "qwen3.8:27b",
"modelComparison": {
"rankings": [
{
"agreementScore": 9,
"assumptionCount": 5,
"confidence": "high",
"model": "qwen3.8:27b",
"score": 6
}
],
"selectedModel": "qwen3.8:27b",
"selectionMethod": "deterministic-consensus-v1",
"selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
},
"patchSha256": "3c6f0376c712f90fde19eba1ca80cd16433ad02587b8237862fd64b60607d8bf",
"patchSummary": "Two one-line changes in the Overmind theme view templates. In app/View/Themed/Overmind/Elements/Events/View/event_general.ctp (line 607) and app/View/Themed/Overmind/Elements/Servers/View/preview_general.ctp (line 323), the raw lbl variable used in the title attribute and inner text of the legend span element is replaced with escapeHtml(lbl). No other logic, routing, or configuration changes are made.",
"patchTruncated": false,
"patches": [
{
"commit": "382188d2f27004118592cf32f970313b31abcb52",
"patchSha256": "3c6f0376c712f90fde19eba1ca80cd16433ad02587b8237862fd64b60607d8bf",
"source": "https://github.com/MISP/misp/commit/382188d2f.patch",
"sourceUrl": "https://github.com/MISP/misp/commit/382188d2f.patch",
"subject": "fix: [security] Escape object names in the Overmind"
}
],
"source": "https://github.com/MISP/misp/commit/382188d2f.patch",
"subject": "fix: [security] Escape object names in the Overmind",
"tagVersionBoundary": {
"commits_after_fix": 55,
"repository": "https://github.com/MISP/misp",
"tag": "v2.5.47",
"version": "2.5.47",
"version_type": "semver"
},
"weaknessRationale": [
{
"cweId": "CWE-79",
"rationale": "User-controlled object names are concatenated directly into an innerHTML string without HTML encoding, allowing injected markup to execute in the victim\u0027s browser. This is a textbook stored XSS via missing output encoding at an HTML sink."
}
]
}
}
},
"recordType": "advisory",
"vulnId": "gcve-1-2026-20206"
}
],
"x_generator": {
"engine": "Vulnogram 0.2.0"
}
}
},
"cveMetadata": {
"assignerOrgId": "00000000-0000-4000-9000-000000000000",
"cveId": "CVE-2026-93296",
"datePublished": "2026-09-17T16:29:00.000Z",
"dateReserved": "2026-09-17T16:29:00.000Z",
"dateUpdated": "2026-09-22T07:36:39.504987Z",
"requesterUserId": "00000000-0000-4000-9000-000000000000",
"serial": 1,
"state": "PUBLISHED",
"vulnId": "gcve-1-2026-20206"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.1"
}
GCVE-1-2026-20043 (CVE-2026-94404)
Vulnerability from gna-1 – Published: 2026-09-21 13:50 – Updated: 2026-09-21 13:50- CWE-352 - Cross-Site Request Forgery (CSRF)
| URL | Tags |
|---|---|
| https://github.com/MISP/MISP/commit/dfbae33f9 | patch |
qwen3.8:27b
advisory
bcp-05-x-01bcp-05-x-02
Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.
| Model | Source | Identifier |
|---|---|---|
| qwen3.8:27b | ollama | qwen3.8:27b |
- Generator
-
patch2vuln.pyon 2026-09-21 13:38 - Model
qwen3.8:27b- Input
-
https://github.com/MISP/MISP/commit/dfbae33f9.patch
beb65b8daaa1… - Confidence
- high
| Commit | Subject | Patch SHA-256 |
|---|---|---|
dfbae33f9a93
|
fix: [security] Keep the CSRF check on the inline attribute | beb65b8daaa1… |
Fix summary
The editField action is removed from the Security component's unlockedActions list and added to the _csrfTokenHeaderOnly array. This enforces CSRF token validation (accepted either in the X-CSRF-Token header or in the request body) while dropping only the field-hash check, which the existing AJAX callers (Overmind index and legacy inline forms) cannot satisfy. Legitimate callers are unaffected because they already transmit a valid CSRF token.
Patch summary
In app/Controller/AttributesController.php, the string 'editField' is added to the _csrfTokenHeaderOnly() call in beforeFilter(), enabling header-or-body CSRF token validation for that action. Simultaneously, the line adding 'editField' to $this->Security->unlockedActions[] is removed, so the action is no longer exempt from CSRF checks. A comment block is added explaining that both the Overmind index (token in X-CSRF-Token header) and legacy inline forms (token in body) already carry a valid token, so only the unsatisfiable field-hash check is dropped.
CVSS rationale
AV:N: The attack is delivered over the network via a malicious web page. AC:L: Crafting a CSRF form is straightforward and requires no race conditions or special timing. AT:N: The target MISP instance is not under the attacker's control. PR:N: The attacker needs no account on MISP; the victim's existing session is sufficient. UI:P: The victim must passively load the attacker's page (e.g., visit a link), which triggers the auto-submitting form; no explicit click or form submission by the victim is required. VC:N: The attack modifies data but does not disclose confidential information. VI:H: The attacker can alter attribute values, types, categories, distribution, and timestamps, corrupting threat-intelligence data for all users. VA:N: No denial-of-service impact is indicated. SC/SI/SA:N: No secondary impacts on other systems are evidenced.
Weakness rationale
- CWE-352 The editField action was placed in unlockedActions, which explicitly bypassed the CSRF token check. This allowed a cross-site form post to modify attribute data using only the victim's session cookie, a textbook CSRF weakness.
Attack pattern rationale
- CAPEC-124 The patch description explicitly states that a cross-site form post carrying the victim's session cookie could rewrite attribute fields. This is the canonical CSRF attack pattern: an attacker induces a victim's browser to send an authenticated request to a target application without the victim's knowledge or consent. CAPEC-124 is the most direct and specific match.
Assumptions to verify
- The affected version range is inferred from the tag_version_boundary (v2.5.47 with 7 commits after the fix); no explicit 'affected' or 'fixed' version strings are provided in the metadata, so the exact version boundary is approximate.
- UI:P (Passive) is assumed because the commit message describes a 'cross-site form post' that auto-submits when the victim's browser loads the attacker's page; if the attack requires the victim to explicitly click a link or submit a form, UI:A would be more appropriate.
- PR:N is assumed because the attacker does not need their own MISP account; the victim's authenticated session cookie is sufficient. If MISP deployments require additional authentication factors (e.g., MFA) that would block a cross-site form post, the effective exploitability may be lower.
- The CAPEC-124 mapping is the closest standard pattern for CSRF; no more specific CAPEC exists for the CakePHP unlockedActions bypass mechanism.
- The Co-Authored-By AI tool is included as a 'tool' credit based on explicit commit metadata; it is not a human contributor.
Model comparison
Selected qwen3.8:27b
by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.
| Model | Score | Agreement | Confidence | Assumptions |
|---|---|---|---|---|
qwen3.8:27b |
6 | 9 | high | 5 |
{
"containers": {
"cna": {
"affected": [
{
"modules": [
"AttributesController"
],
"product": "MISP",
"programFiles": [
"app/Controller/AttributesController.php"
],
"repo": "https://github.com/MISP/MISP",
"vendor": "MISP",
"versions": [
{
"lessThan": "2.5.47",
"status": "affected",
"version": "0",
"versionType": "semver"
}
]
}
],
"credits": [
{
"lang": "en",
"type": "reporter",
"value": "Jeroen Pinoy"
},
{
"lang": "en",
"type": "remediation developer",
"value": "iglocska"
},
{
"lang": "en",
"type": "remediation developer",
"value": "Claude Opus 5 (1M context)"
}
],
"descriptions": [
{
"lang": "en",
"supportingMedia": [
{
"base64": false,
"type": "text/html",
"value": "\u003cp\u003eMISP has a security issue that could let an attacker change threat-intelligence data through a logged-in user\u2019s browser without that user knowingly approving the change.\u003c/p\u003e\u003cp\u003eThe affected function did not properly enforce MISP\u2019s usual protection against forged requests. Because of this, an attacker could create a malicious webpage that silently sends a request to MISP when visited by an authenticated user.\u003c/p\u003e\u003cp\u003eIf successful, the attacker could change details of an attribute, such as its value, type, category, comment, distribution settings, or related timestamps.\u003c/p\u003e\u003cp\u003eThe attack requires the victim to already be logged in to MISP and to visit an attacker-controlled page.\u003c/p\u003e\u003cp\u003eThe main impact is unauthorized modification of threat-intelligence data, which could lead to incorrect indicators, wrong classifications, or altered sharing settings and reduce confidence in the accuracy of the information stored in MISP.\u003c/p\u003eVersion affected: \u0026lt;2.5.47"
}
],
"value": "MISP has a security issue that could let an attacker change threat-intelligence data through a logged-in user\u2019s browser without that user knowingly approving the change.\n\nThe affected function did not properly enforce MISP\u2019s usual protection against forged requests. Because of this, an attacker could create a malicious webpage that silently sends a request to MISP when visited by an authenticated user.\n\nIf successful, the attacker could change details of an attribute, such as its value, type, category, comment, distribution settings, or related timestamps.\n\nThe attack requires the victim to already be logged in to MISP and to visit an attacker-controlled page.\n\nThe main impact is unauthorized modification of threat-intelligence data, which could lead to incorrect indicators, wrong classifications, or altered sharing settings and reduce confidence in the accuracy of the information stored in MISP.\n\nVersion affected: \u003c2.5.47"
}
],
"impacts": [
{
"capecId": "CAPEC-124",
"descriptions": [
{
"lang": "en",
"value": "CAPEC-124 Cross-Site Request Forgery"
}
]
}
],
"metrics": [
{
"cvssV4_0": {
"Automatable": "NOT_DEFINED",
"Recovery": "NOT_DEFINED",
"Safety": "NOT_DEFINED",
"attackComplexity": "LOW",
"attackRequirements": "NONE",
"attackVector": "NETWORK",
"baseScore": 7.1,
"baseSeverity": "HIGH",
"privilegesRequired": "NONE",
"providerUrgency": "NOT_DEFINED",
"subAvailabilityImpact": "NONE",
"subConfidentialityImpact": "NONE",
"subIntegrityImpact": "NONE",
"userInteraction": "PASSIVE",
"valueDensity": "NOT_DEFINED",
"vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:N/VI:H/VA:N/SC:N/SI:N/SA:N",
"version": "4.0",
"vulnAvailabilityImpact": "NONE",
"vulnConfidentialityImpact": "NONE",
"vulnIntegrityImpact": "HIGH",
"vulnerabilityResponseEffort": "NOT_DEFINED"
},
"format": "CVSS",
"scenarios": [
{
"lang": "en",
"value": "GENERAL"
}
]
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-352",
"description": "CWE-352 Cross-Site Request Forgery (CSRF)",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"orgId": "00000000-0000-4000-9000-000000000000"
},
"references": [
{
"name": "Security patch",
"tags": [
"patch"
],
"url": "https://github.com/MISP/MISP/commit/dfbae33f9"
}
],
"solutions": [
{
"lang": "en",
"supportingMedia": [
{
"base64": false,
"type": "text/html",
"value": "\u003cp\u003eThe editField action is removed from the Security component\u0027s unlockedActions list and added to the _csrfTokenHeaderOnly array. This enforces CSRF token validation (accepted either in the X-CSRF-Token header or in the request body) while dropping only the field-hash check, which the existing AJAX callers (Overmind index and legacy inline forms) cannot satisfy. Legitimate callers are unaffected because they already transmit a valid CSRF token.\u003c/p\u003e"
}
],
"value": "The editField action is removed from the Security component\u0027s unlockedActions list and added to the _csrfTokenHeaderOnly array. This enforces CSRF token validation (accepted either in the X-CSRF-Token header or in the request body) while dropping only the field-hash check, which the existing AJAX callers (Overmind index and legacy inline forms) cannot satisfy. Legitimate callers are unaffected because they already transmit a valid CSRF token."
}
],
"title": "MISP CSRF vulnerability allows unauthorized attribute modification",
"x_gcve": [
{
"extensions": {
"bcp-05-x-01": {
"ai_annotations": [
{
"ai_level": "generated",
"description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
"gna_source": 1,
"models": [
{
"gna_source": 1,
"identifier": "qwen3.8:27b",
"name": "qwen3.8:27b",
"source": "ollama"
}
],
"review_status": "review",
"scope": "record",
"tags": [
"ai-computer-assisted:llm-generated",
"ai-computer-assisted:classification"
]
}
]
},
"bcp-05-x-02": {
"x_patch2vuln": {
"assumptions": [
"The affected version range is inferred from the tag_version_boundary (v2.5.47 with 7 commits after the fix); no explicit \u0027affected\u0027 or \u0027fixed\u0027 version strings are provided in the metadata, so the exact version boundary is approximate.",
"UI:P (Passive) is assumed because the commit message describes a \u0027cross-site form post\u0027 that auto-submits when the victim\u0027s browser loads the attacker\u0027s page; if the attack requires the victim to explicitly click a link or submit a form, UI:A would be more appropriate.",
"PR:N is assumed because the attacker does not need their own MISP account; the victim\u0027s authenticated session cookie is sufficient. If MISP deployments require additional authentication factors (e.g., MFA) that would block a cross-site form post, the effective exploitability may be lower.",
"The CAPEC-124 mapping is the closest standard pattern for CSRF; no more specific CAPEC exists for the CakePHP unlockedActions bypass mechanism.",
"The Co-Authored-By AI tool is included as a \u0027tool\u0027 credit based on explicit commit metadata; it is not a human contributor."
],
"capecRationale": [
{
"capecId": "CAPEC-124",
"rationale": "The patch description explicitly states that a cross-site form post carrying the victim\u0027s session cookie could rewrite attribute fields. This is the canonical CSRF attack pattern: an attacker induces a victim\u0027s browser to send an authenticated request to a target application without the victim\u0027s knowledge or consent. CAPEC-124 is the most direct and specific match."
}
],
"commit": "dfbae33f9a93bf127fae115bc6cd1a9e03d028e9",
"confidence": "high",
"credits": [
{
"lang": "en",
"type": "reporter",
"value": "Jeroen Pinoy"
},
{
"lang": "en",
"type": "remediation developer",
"value": "iglocska"
},
{
"lang": "en",
"type": "remediation developer",
"value": "Claude Opus 5 (1M context)"
}
],
"cvssRationale": "AV:N: The attack is delivered over the network via a malicious web page. AC:L: Crafting a CSRF form is straightforward and requires no race conditions or special timing. AT:N: The target MISP instance is not under the attacker\u0027s control. PR:N: The attacker needs no account on MISP; the victim\u0027s existing session is sufficient. UI:P: The victim must passively load the attacker\u0027s page (e.g., visit a link), which triggers the auto-submitting form; no explicit click or form submission by the victim is required. VC:N: The attack modifies data but does not disclose confidential information. VI:H: The attacker can alter attribute values, types, categories, distribution, and timestamps, corrupting threat-intelligence data for all users. VA:N: No denial-of-service impact is indicated. SC/SI/SA:N: No secondary impacts on other systems are evidenced.",
"fixSummary": "The editField action is removed from the Security component\u0027s unlockedActions list and added to the _csrfTokenHeaderOnly array. This enforces CSRF token validation (accepted either in the X-CSRF-Token header or in the request body) while dropping only the field-hash check, which the existing AJAX callers (Overmind index and legacy inline forms) cannot satisfy. Legitimate callers are unaffected because they already transmit a valid CSRF token.",
"generatedAt": "2026-09-21T13:38:54.473662Z",
"generator": "patch2vuln.py",
"model": "qwen3.8:27b",
"modelComparison": {
"rankings": [
{
"agreementScore": 9,
"assumptionCount": 5,
"confidence": "high",
"model": "qwen3.8:27b",
"score": 6
}
],
"selectedModel": "qwen3.8:27b",
"selectionMethod": "deterministic-consensus-v1",
"selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
},
"patchSha256": "beb65b8daaa1d2e649f1178cc7f2c774f4ba7e75a13cc000524eef566c4545c8",
"patchSummary": "In app/Controller/AttributesController.php, the string \u0027editField\u0027 is added to the _csrfTokenHeaderOnly() call in beforeFilter(), enabling header-or-body CSRF token validation for that action. Simultaneously, the line adding \u0027editField\u0027 to $this-\u003eSecurity-\u003eunlockedActions[] is removed, so the action is no longer exempt from CSRF checks. A comment block is added explaining that both the Overmind index (token in X-CSRF-Token header) and legacy inline forms (token in body) already carry a valid token, so only the unsatisfiable field-hash check is dropped.",
"patchTruncated": false,
"patches": [
{
"commit": "dfbae33f9a93bf127fae115bc6cd1a9e03d028e9",
"patchSha256": "beb65b8daaa1d2e649f1178cc7f2c774f4ba7e75a13cc000524eef566c4545c8",
"source": "https://github.com/MISP/MISP/commit/dfbae33f9.patch",
"sourceUrl": "https://github.com/MISP/MISP/commit/dfbae33f9.patch",
"subject": "fix: [security] Keep the CSRF check on the inline attribute"
}
],
"source": "https://github.com/MISP/MISP/commit/dfbae33f9.patch",
"subject": "fix: [security] Keep the CSRF check on the inline attribute",
"tagVersionBoundary": {
"commits_after_fix": 7,
"repository": "https://github.com/MISP/MISP",
"tag": "v2.5.47",
"version": "2.5.47",
"version_type": "semver"
},
"weaknessRationale": [
{
"cweId": "CWE-352",
"rationale": "The editField action was placed in unlockedActions, which explicitly bypassed the CSRF token check. This allowed a cross-site form post to modify attribute data using only the victim\u0027s session cookie, a textbook CSRF weakness."
}
]
}
}
},
"recordType": "advisory",
"vulnId": "GCVE-1-2026-20043"
}
]
}
},
"cveMetadata": {
"assignerOrgId": "00000000-0000-4000-9000-000000000000",
"cveId": "CVE-2026-94404",
"datePublished": "2026-09-21T13:50:37.643790Z",
"dateReserved": "2026-09-21T13:50:44.676Z",
"dateUpdated": "2026-09-21T13:50:44.744216Z",
"state": "PUBLISHED",
"vulnId": "GCVE-1-2026-20043"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.1"
}
GCVE-1-2026-20056 (CVE-2026-94401)
Vulnerability from gna-1 – Published: 2026-09-21 13:35 – Updated: 2026-09-21 13:36| URL | Tags |
|---|---|
| https://github.com/MISP/MISP/commit/d5f247b91 | patch |
qwen3.8:27b
advisory
bcp-05-x-01bcp-05-x-02
Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.
| Model | Source | Identifier |
|---|---|---|
| qwen3.8:27b | ollama | qwen3.8:27b |
- Generator
-
patch2vuln.pyon 2026-09-21 13:28 - Model
qwen3.8:27b- Input
-
https://github.com/MISP/MISP/commit/d5f247b91.patch
004f65b4d416… - Confidence
- high
| Commit | Subject | Patch SHA-256 |
|---|---|---|
d5f247b91ad3
|
fix: [security] Refuse a MISP export upload whose content is | 004f65b4d416… |
Fix summary
The fix introduces two defensive measures in addMISPExportFile(): (1) a pre-parse validation that rejects any uploaded content not containing an XML document marker (the '<' character), preventing path or URL strings from reaching the XML parser; and (2) an explicit readFile => false option passed to Xml::build(), which disables the library's ability to interpret the input as a file path or URL to be read or fetched. Together these changes ensure that only genuine XML document content is parsed and that no server-side file access or network request is triggered by user-supplied data.
Patch summary
In app/Model/Event.php, the single line calling Xml::toArray(Xml::build($data)) is replaced with a guard clause that throws an Exception if strpos($data, '<') === false, followed by Xml::toArray(Xml::build($data, ['readFile' => false])). Nine lines are added (including a four-line explanatory comment) and one line is removed. No other files are modified.
CVSS rationale
AV:N: exploitation occurs over the network via the MISP web interface. AC:L: the attack is deterministic; uploading a file with a path or URL as content reliably triggers the read/fetch. AT:N: no special target-side conditions are required. PR:L: the attacker needs an authenticated account with perm_modify (a non-privileged but authenticated role). UI:N: the attacker performs the upload autonomously; no victim interaction is needed. VC:H: arbitrary local file read exposes sensitive server data (credentials, configs, keys). VI:N: the imported data is stored as an event but does not corrupt application state or logic. VA:N: no denial-of-service impact is indicated. SC:H: the SSRF component allows the attacker to read responses from internal services (metadata endpoints, admin panels, etc.), impacting subsequent-component confidentiality. SI:N and SA:N: no evidence of integrity or availability impact on subsequent components from the patch evidence.
Weakness rationale
- CWE-918 The primary and most severe impact is that an attacker-supplied URL in the uploaded file body causes the MISP server to issue an HTTP request to an arbitrary internal or external address and import the response. This is a textbook SSRF: user-controlled input is used as a network request target by the server.
- CWE-73 The secondary impact is that a user-supplied local file path in the uploaded file body is passed to Xml::build() which reads that file from the server's filesystem. The attacker controls the file path that the server reads, matching CWE-73. CWE-22 (Path Traversal) was considered but the input is not a relative traversal sequence; it is an absolute path chosen by the attacker, making CWE-73 the more precise mapping.
Attack pattern rationale
- CAPEC-126 The patch explicitly describes the SSRF vector: an .xml upload containing a URL causes MISP to fetch that URL and import the result. CAPEC-126 describes an attacker manipulating a server-side application to make it issue requests to unintended targets, which is exactly the attack demonstrated here. The local-file-read aspect is a secondary manifestation of the same root cause (unvalidated input passed to a function that resolves locators) and is subsumed by the SSRF pattern in this context.
- CAPEC-2 The attacker manipulates the content parameter of the MISP export upload to supply a file path or URL instead of a valid XML document. CAPEC-2 covers the general technique of altering application parameters to trigger unintended behavior. This is a secondary mapping; CAPEC-126 is the more specific and primary pattern. The uncertainty is that CAPEC-2 is broader and less precise than CAPEC-126 for this specific vulnerability.
Assumptions to verify
- The commit message states the vulnerability was 'found during the internal review, not externally reported,' so no external finder or reporter credit is assigned. The internal reviewer's identity is not specified in the patch metadata.
- The affected version range is inferred from the tag_version_boundary (v2.5.47 with 40 commits after the fix), but exact affected and fixed version numbers are not stated in the patch. The boundary is recorded as a reference point only.
- PR:L assumes that perm_modify is a non-administrative, commonly granted role in MISP deployments. If perm_modify is equivalent to full administrative access in some configurations, PR could be elevated to PR:H.
- The CAPEC-2 mapping is a broader, secondary pattern included for completeness; CAPEC-126 is the primary and most specific match. The uncertainty is noted in the CAPEC-2 rationale.
- The commit message references an operator-precedence issue in Xml::build()'s readFile guard not covering the https branch. This is taken as evidence from the commit message and patch comments; the actual Xml utility source is not included in the patch, so the exact mechanism of the https bypass is assumed rather than verified.
- CVSS SC:H assumes the SSRF can reach internal services with sensitive data. If the deployment restricts outbound network access, SC could be reduced.
Model comparison
Selected qwen3.8:27b
by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.
| Model | Score | Agreement | Confidence | Assumptions |
|---|---|---|---|---|
qwen3.8:27b |
5 | 9 | high | 6 |
{
"containers": {
"cna": {
"affected": [
{
"modules": [
"app/Model/Event.php (addMISPExportFile)"
],
"product": "MISP",
"programFiles": [
"app/Model/Event.php"
],
"repo": "https://github.com/MISP/MISP",
"vendor": "MISP",
"versions": [
{
"lessThan": "2.5.47",
"status": "affected",
"version": "0",
"versionType": "semver"
}
]
}
],
"credits": [
{
"lang": "en",
"type": "remediation developer",
"value": "iglocska"
},
{
"lang": "en",
"type": "remediation developer",
"value": "Claude Opus 5 (1M context)"
}
],
"descriptions": [
{
"lang": "en",
"supportingMedia": [
{
"base64": false,
"type": "text/html",
"value": "\u003cp\u003eMISP has a file-handling vulnerability that could let certain authenticated users make the server read files or access internal network services.\u003c/p\u003e\u003cp\u003eWhen importing an XML file, MISP did not properly verify that the uploaded content was actually XML. Because of this, a user with permission to modify data could upload a file containing a local file path or a web address instead.\u003c/p\u003e\u003cp\u003eIf a local file path was supplied, MISP could read that file from the server. If a URL was supplied, MISP could make a request to that address, including systems that may only be reachable from inside the organization\u2019s network.\u003c/p\u003e\u003cp\u003eThe vulnerability could therefore expose sensitive local files and allow unauthorized requests to internal services.\u003c/p\u003e\u003cp\u003eExploitation required a valid MISP account with modify permissions, but no additional user interaction was needed.\u003c/p\u003eVersion affected: \u0026lt;2.5.47"
}
],
"value": "MISP has a file-handling vulnerability that could let certain authenticated users make the server read files or access internal network services.\n\nWhen importing an XML file, MISP did not properly verify that the uploaded content was actually XML. Because of this, a user with permission to modify data could upload a file containing a local file path or a web address instead.\n\nIf a local file path was supplied, MISP could read that file from the server. If a URL was supplied, MISP could make a request to that address, including systems that may only be reachable from inside the organization\u2019s network.\n\nThe vulnerability could therefore expose sensitive local files and allow unauthorized requests to internal services.\n\nExploitation required a valid MISP account with modify permissions, but no additional user interaction was needed.\n\nVersion affected: \u003c2.5.47"
}
],
"impacts": [
{
"capecId": "CAPEC-126",
"descriptions": [
{
"lang": "en",
"value": "CAPEC-126 Server Side Request Forgery"
}
]
},
{
"capecId": "CAPEC-2",
"descriptions": [
{
"lang": "en",
"value": "CAPEC-2 Parameter Manipulation"
}
]
}
],
"metrics": [
{
"cvssV4_0": {
"Automatable": "NOT_DEFINED",
"Recovery": "NOT_DEFINED",
"Safety": "NOT_DEFINED",
"attackComplexity": "LOW",
"attackRequirements": "NONE",
"attackVector": "NETWORK",
"baseScore": 8.3,
"baseSeverity": "HIGH",
"privilegesRequired": "LOW",
"providerUrgency": "NOT_DEFINED",
"subAvailabilityImpact": "NONE",
"subConfidentialityImpact": "HIGH",
"subIntegrityImpact": "NONE",
"userInteraction": "NONE",
"valueDensity": "NOT_DEFINED",
"vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:H/VI:N/VA:N/SC:H/SI:N/SA:N",
"version": "4.0",
"vulnAvailabilityImpact": "NONE",
"vulnConfidentialityImpact": "HIGH",
"vulnIntegrityImpact": "NONE",
"vulnerabilityResponseEffort": "NOT_DEFINED"
},
"format": "CVSS",
"scenarios": [
{
"lang": "en",
"value": "GENERAL"
}
]
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-918",
"description": "CWE-918 Server-Side Request Forgery (SSRF)",
"lang": "en",
"type": "CWE"
}
]
},
{
"descriptions": [
{
"cweId": "CWE-73",
"description": "CWE-73 External Control of File Name or Path",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"orgId": "00000000-0000-4000-9000-000000000000"
},
"references": [
{
"name": "Security patch",
"tags": [
"patch"
],
"url": "https://github.com/MISP/MISP/commit/d5f247b91"
}
],
"solutions": [
{
"lang": "en",
"supportingMedia": [
{
"base64": false,
"type": "text/html",
"value": "\u003cp\u003eThe fix introduces two defensive measures in addMISPExportFile(): (1) a pre-parse validation that rejects any uploaded content not containing an XML document marker (the \u0027\u0026lt;\u0027 character), preventing path or URL strings from reaching the XML parser; and (2) an explicit readFile =\u0026gt; false option passed to Xml::build(), which disables the library\u0027s ability to interpret the input as a file path or URL to be read or fetched. Together these changes ensure that only genuine XML document content is parsed and that no server-side file access or network request is triggered by user-supplied data.\u003c/p\u003e"
}
],
"value": "The fix introduces two defensive measures in addMISPExportFile(): (1) a pre-parse validation that rejects any uploaded content not containing an XML document marker (the \u0027\u003c\u0027 character), preventing path or URL strings from reaching the XML parser; and (2) an explicit readFile =\u003e false option passed to Xml::build(), which disables the library\u0027s ability to interpret the input as a file path or URL to be read or fetched. Together these changes ensure that only genuine XML document content is parsed and that no server-side file access or network request is triggered by user-supplied data."
}
],
"title": "MISP Arbitrary Local File Read and SSRF via MISP Export Upload",
"x_gcve": [
{
"extensions": {
"bcp-05-x-01": {
"ai_annotations": [
{
"ai_level": "generated",
"description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
"gna_source": 1,
"models": [
{
"gna_source": 1,
"identifier": "qwen3.8:27b",
"name": "qwen3.8:27b",
"source": "ollama"
}
],
"review_status": "review",
"scope": "record",
"tags": [
"ai-computer-assisted:llm-generated",
"ai-computer-assisted:classification"
]
}
]
},
"bcp-05-x-02": {
"x_patch2vuln": {
"assumptions": [
"The commit message states the vulnerability was \u0027found during the internal review, not externally reported,\u0027 so no external finder or reporter credit is assigned. The internal reviewer\u0027s identity is not specified in the patch metadata.",
"The affected version range is inferred from the tag_version_boundary (v2.5.47 with 40 commits after the fix), but exact affected and fixed version numbers are not stated in the patch. The boundary is recorded as a reference point only.",
"PR:L assumes that perm_modify is a non-administrative, commonly granted role in MISP deployments. If perm_modify is equivalent to full administrative access in some configurations, PR could be elevated to PR:H.",
"The CAPEC-2 mapping is a broader, secondary pattern included for completeness; CAPEC-126 is the primary and most specific match. The uncertainty is noted in the CAPEC-2 rationale.",
"The commit message references an operator-precedence issue in Xml::build()\u0027s readFile guard not covering the https branch. This is taken as evidence from the commit message and patch comments; the actual Xml utility source is not included in the patch, so the exact mechanism of the https bypass is assumed rather than verified.",
"CVSS SC:H assumes the SSRF can reach internal services with sensitive data. If the deployment restricts outbound network access, SC could be reduced."
],
"capecRationale": [
{
"capecId": "CAPEC-126",
"rationale": "The patch explicitly describes the SSRF vector: an .xml upload containing a URL causes MISP to fetch that URL and import the result. CAPEC-126 describes an attacker manipulating a server-side application to make it issue requests to unintended targets, which is exactly the attack demonstrated here. The local-file-read aspect is a secondary manifestation of the same root cause (unvalidated input passed to a function that resolves locators) and is subsumed by the SSRF pattern in this context."
},
{
"capecId": "CAPEC-2",
"rationale": "The attacker manipulates the content parameter of the MISP export upload to supply a file path or URL instead of a valid XML document. CAPEC-2 covers the general technique of altering application parameters to trigger unintended behavior. This is a secondary mapping; CAPEC-126 is the more specific and primary pattern. The uncertainty is that CAPEC-2 is broader and less precise than CAPEC-126 for this specific vulnerability."
}
],
"commit": "d5f247b91ad3e8ea86feaaa587bf0466c95ea996",
"confidence": "high",
"credits": [
{
"lang": "en",
"type": "remediation developer",
"value": "iglocska"
},
{
"lang": "en",
"type": "remediation developer",
"value": "Claude Opus 5 (1M context)"
}
],
"cvssRationale": "AV:N: exploitation occurs over the network via the MISP web interface. AC:L: the attack is deterministic; uploading a file with a path or URL as content reliably triggers the read/fetch. AT:N: no special target-side conditions are required. PR:L: the attacker needs an authenticated account with perm_modify (a non-privileged but authenticated role). UI:N: the attacker performs the upload autonomously; no victim interaction is needed. VC:H: arbitrary local file read exposes sensitive server data (credentials, configs, keys). VI:N: the imported data is stored as an event but does not corrupt application state or logic. VA:N: no denial-of-service impact is indicated. SC:H: the SSRF component allows the attacker to read responses from internal services (metadata endpoints, admin panels, etc.), impacting subsequent-component confidentiality. SI:N and SA:N: no evidence of integrity or availability impact on subsequent components from the patch evidence.",
"fixSummary": "The fix introduces two defensive measures in addMISPExportFile(): (1) a pre-parse validation that rejects any uploaded content not containing an XML document marker (the \u0027\u003c\u0027 character), preventing path or URL strings from reaching the XML parser; and (2) an explicit readFile =\u003e false option passed to Xml::build(), which disables the library\u0027s ability to interpret the input as a file path or URL to be read or fetched. Together these changes ensure that only genuine XML document content is parsed and that no server-side file access or network request is triggered by user-supplied data.",
"generatedAt": "2026-09-21T13:28:43.719596Z",
"generator": "patch2vuln.py",
"model": "qwen3.8:27b",
"modelComparison": {
"rankings": [
{
"agreementScore": 9,
"assumptionCount": 6,
"confidence": "high",
"model": "qwen3.8:27b",
"score": 5
}
],
"selectedModel": "qwen3.8:27b",
"selectionMethod": "deterministic-consensus-v1",
"selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
},
"patchSha256": "004f65b4d416b8e01f8b713a982a9b4d808e454c40cd74cc94d36137eb1b2845",
"patchSummary": "In app/Model/Event.php, the single line calling Xml::toArray(Xml::build($data)) is replaced with a guard clause that throws an Exception if strpos($data, \u0027\u003c\u0027) === false, followed by Xml::toArray(Xml::build($data, [\u0027readFile\u0027 =\u003e false])). Nine lines are added (including a four-line explanatory comment) and one line is removed. No other files are modified.",
"patchTruncated": false,
"patches": [
{
"commit": "d5f247b91ad3e8ea86feaaa587bf0466c95ea996",
"patchSha256": "004f65b4d416b8e01f8b713a982a9b4d808e454c40cd74cc94d36137eb1b2845",
"source": "https://github.com/MISP/MISP/commit/d5f247b91.patch",
"sourceUrl": "https://github.com/MISP/MISP/commit/d5f247b91.patch",
"subject": "fix: [security] Refuse a MISP export upload whose content is"
}
],
"source": "https://github.com/MISP/MISP/commit/d5f247b91.patch",
"subject": "fix: [security] Refuse a MISP export upload whose content is",
"tagVersionBoundary": {
"commits_after_fix": 40,
"repository": "https://github.com/MISP/MISP",
"tag": "v2.5.47",
"version": "2.5.47",
"version_type": "semver"
},
"weaknessRationale": [
{
"cweId": "CWE-918",
"rationale": "The primary and most severe impact is that an attacker-supplied URL in the uploaded file body causes the MISP server to issue an HTTP request to an arbitrary internal or external address and import the response. This is a textbook SSRF: user-controlled input is used as a network request target by the server."
},
{
"cweId": "CWE-73",
"rationale": "The secondary impact is that a user-supplied local file path in the uploaded file body is passed to Xml::build() which reads that file from the server\u0027s filesystem. The attacker controls the file path that the server reads, matching CWE-73. CWE-22 (Path Traversal) was considered but the input is not a relative traversal sequence; it is an absolute path chosen by the attacker, making CWE-73 the more precise mapping."
}
]
}
}
},
"recordType": "advisory",
"vulnId": "GCVE-1-2026-20056"
}
]
}
},
"cveMetadata": {
"assignerOrgId": "00000000-0000-4000-9000-000000000000",
"cveId": "CVE-2026-94401",
"datePublished": "2026-09-21T13:35:58.783318Z",
"dateReserved": "2026-09-21T13:36:11.048Z",
"dateUpdated": "2026-09-21T13:36:11.123285Z",
"state": "PUBLISHED",
"vulnId": "GCVE-1-2026-20056"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.1"
}
GCVE-1-2026-20106 (CVE-2026-94394)
Vulnerability from gna-1 – Published: 2026-09-21 13:25 – Updated: 2026-09-21 13:25- CWE-862 - Missing Authorization
| URL | Tags |
|---|---|
| https://github.com/MISP/MISP/commit/f36634e57 | patch |
qwen3.8:27b
advisory
bcp-05-x-01bcp-05-x-02
Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.
| Model | Source | Identifier |
|---|---|---|
| qwen3.8:27b | ollama | qwen3.8:27b |
- Generator
-
patch2vuln.pyon 2026-09-21 13:15 - Model
qwen3.8:27b- Input
-
https://github.com/MISP/MISP/commit/f36634e57.patch
2688fca1334e… - Confidence
- medium
| Commit | Subject | Patch SHA-256 |
|---|---|---|
f36634e57b93
|
fix: [security] Extended event granular restrictions ignored | 2688fca1334e… |
Fix summary
The fix introduces granular authorization checks in the ObjectReferencesController add() method. For non-site-admin users, the query conditions for attributes, objects, and object attributes are now augmented with OR clauses that restrict results to: (1) entities belonging to the user's own event, (2) entities with unrestricted distribution levels (1, 2, 3, 5), or (3) entities with distribution level 4 whose sharing_group_id is in the user's authorized sharing group list. This ensures that only data the user is explicitly authorized to see under MISP's distribution and sharing-group model is included in the object reference operation.
Patch summary
In app/Controller/ObjectReferencesController.php, the add() method is modified to build dynamic condition arrays ($attributeConditions, $objectConditions, $objectAttributeConditions) that incorporate the current user's role and authorized sharing group IDs. For non-site-admin users, each condition array gains an OR clause filtering by event ownership, distribution levels (1,2,3,5), or distribution level 4 with a matching sharing_group_id from the user's authorized IDs. These condition arrays replace the previously hardcoded minimal conditions (deleted=0, object_id=0) in the Contain clauses for Attribute, Object, and nested Object.Attribute queries. The user's authorized sharing group IDs are obtained via SharingGroup->authorizedIds($user).
CVSS rationale
AV:N: MISP is a network-accessible web application. AC:L: The attack requires no special conditions beyond having a valid account with event-level access; no race conditions or complex setup needed. AT:N: No manipulation of a separate attack target is required. PR:L: The attacker needs a low-privilege authenticated account (any non-site-admin user with access to an event containing restricted attributes/objects). UI:N: No victim interaction is required; the attacker simply calls the API endpoint. VC:N/VI:N/VA:N: The MISP server itself (the vulnerable component) is not compromised in its own confidentiality, integrity, or availability. SC:H: The resource impact is high because restricted attribute values, object metadata, and object-attribute data that are explicitly gated by distribution levels and sharing groups are exposed to an unauthorized user. SI:N: No integrity impact on the data. SA:N: No safety impact.
Weakness rationale
- CWE-862 The ObjectReferencesController add() method was missing authorization checks for granular distribution-level and sharing-group restrictions on attributes, objects, and object attributes. Only event-level access was verified, while the sub-entity level access controls (distribution levels 1-5, sharing group membership) were entirely absent from the query conditions. This is a classic missing authorization check at a finer granularity than what was enforced.
Attack pattern rationale
- CAPEC-109 The closest plausible CAPEC is Parameter Tampering. An authenticated user supplies an objectId parameter to the add() endpoint, and the application fails to enforce proper authorization on the sub-entities (attributes, objects, object attributes) associated with that object. The user effectively 'tampers' with the scope of accessible data by leveraging event-level access to reach restricted sub-entities. This is not a perfect match because the user is not modifying a parameter to an invalid value but rather exploiting the absence of a check on a valid parameter's sub-resources. No CAPEC specifically models 'broken granular access control' or 'insecure direct object reference at sub-entity level,' making CAPEC-109 the best available approximation.
Assumptions to verify
- The exact affected and fixed version numbers are not specified in the patch metadata. The tag_version_boundary indicates the fix commit is 140 commits after v2.5.47, but the precise release version containing the fix is unknown.
- The CAPEC-109 (Parameter Tampering) mapping is the closest available match but is not a perfect fit; the vulnerability is more precisely a missing granular authorization check rather than parameter manipulation. No CAPEC specifically models broken sub-entity access control.
- The CVSS SC:H rating assumes that the exposed attribute/object data constitutes sensitive information (e.g., IOCs, malware indicators, threat intelligence) whose unauthorized disclosure is a significant confidentiality impact. If the data is considered low-sensitivity, SC:L may be more appropriate.
- The vulnerability requires the attacker to already have event-level access; it does not grant access to events the user cannot see. The bypass is limited to granular restrictions within an already-accessible event.
- The patch only addresses the add() method in ObjectReferencesController; other methods or controllers with similar patterns may or may not be affected, but no evidence of additional affected code paths is present in this patch.
Model comparison
Selected qwen3.8:27b
by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.
| Model | Score | Agreement | Confidence | Assumptions |
|---|---|---|---|---|
qwen3.8:27b |
5 | 9 | medium | 5 |
{
"containers": {
"cna": {
"affected": [
{
"modules": [
"ObjectReferencesController"
],
"product": "MISP",
"programFiles": [
"app/Controller/ObjectReferencesController.php"
],
"repo": "https://github.com/MISP/MISP",
"vendor": "MISP",
"versions": [
{
"lessThan": "2.5.47",
"status": "affected",
"version": "0",
"versionType": "semver"
}
]
}
],
"credits": [
{
"lang": "en",
"type": "reporter",
"value": "Jeroen Pinoy"
},
{
"lang": "en",
"type": "remediation developer",
"value": "iglocska"
}
],
"descriptions": [
{
"lang": "en",
"supportingMedia": [
{
"base64": false,
"type": "text/html",
"value": "When a regular user adds a reference between objects or attributes, MISP checks whether the user can access the overall event, but it does not always check whether the individual pieces of data are also allowed for that user.\u003cbr\u003e\u003cbr\u003e\u003cp\u003eBecause of this, someone who can view an event could potentially access attributes or objects inside that event that were meant to be restricted to a specific sharing group or distribution level.\u003c/p\u003eThe vulnerability affects authenticated users who are not site administrators and who already have access to an event containing more restricted data.\u003cbr\u003e\u003cbr\u003eThe main impact is that users may be able to view sensitive attribute values, object details, or related object data that they should not normally be allowed to see.\u003cbr\u003e"
}
],
"value": "When a regular user adds a reference between objects or attributes, MISP checks whether the user can access the overall event, but it does not always check whether the individual pieces of data are also allowed for that user.\n\nBecause of this, someone who can view an event could potentially access attributes or objects inside that event that were meant to be restricted to a specific sharing group or distribution level.\n\nThe vulnerability affects authenticated users who are not site administrators and who already have access to an event containing more restricted data.\n\nThe main impact is that users may be able to view sensitive attribute values, object details, or related object data that they should not normally be allowed to see."
}
],
"impacts": [
{
"capecId": "CAPEC-109",
"descriptions": [
{
"lang": "en",
"value": "CAPEC-109 Parameter Tampering"
}
]
}
],
"metrics": [
{
"cvssV4_0": {
"Automatable": "NOT_DEFINED",
"Recovery": "NOT_DEFINED",
"Safety": "NOT_DEFINED",
"attackComplexity": "LOW",
"attackRequirements": "NONE",
"attackVector": "NETWORK",
"baseScore": 6.3,
"baseSeverity": "MEDIUM",
"privilegesRequired": "LOW",
"providerUrgency": "NOT_DEFINED",
"subAvailabilityImpact": "NONE",
"subConfidentialityImpact": "HIGH",
"subIntegrityImpact": "NONE",
"userInteraction": "NONE",
"valueDensity": "NOT_DEFINED",
"vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:N/VI:N/VA:N/SC:H/SI:N/SA:N",
"version": "4.0",
"vulnAvailabilityImpact": "NONE",
"vulnConfidentialityImpact": "NONE",
"vulnIntegrityImpact": "NONE",
"vulnerabilityResponseEffort": "NOT_DEFINED"
},
"format": "CVSS",
"scenarios": [
{
"lang": "en",
"value": "GENERAL"
}
]
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-862",
"description": "CWE-862 Missing Authorization",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"orgId": "00000000-0000-4000-9000-000000000000"
},
"references": [
{
"name": "Security patch",
"tags": [
"patch"
],
"url": "https://github.com/MISP/MISP/commit/f36634e57"
}
],
"solutions": [
{
"lang": "en",
"supportingMedia": [
{
"base64": false,
"type": "text/html",
"value": "\u003cp\u003eThe fix introduces granular authorization checks in the ObjectReferencesController add() method. For non-site-admin users, the query conditions for attributes, objects, and object attributes are now augmented with OR clauses that restrict results to: (1) entities belonging to the user\u0027s own event, (2) entities with unrestricted distribution levels (1, 2, 3, 5), or (3) entities with distribution level 4 whose sharing_group_id is in the user\u0027s authorized sharing group list. This ensures that only data the user is explicitly authorized to see under MISP\u0027s distribution and sharing-group model is included in the object reference operation.\u003c/p\u003e"
}
],
"value": "The fix introduces granular authorization checks in the ObjectReferencesController add() method. For non-site-admin users, the query conditions for attributes, objects, and object attributes are now augmented with OR clauses that restrict results to: (1) entities belonging to the user\u0027s own event, (2) entities with unrestricted distribution levels (1, 2, 3, 5), or (3) entities with distribution level 4 whose sharing_group_id is in the user\u0027s authorized sharing group list. This ensures that only data the user is explicitly authorized to see under MISP\u0027s distribution and sharing-group model is included in the object reference operation."
}
],
"title": "MISP ObjectReferencesController: Granular Distribution and Sharing Group Restrictions Bypassed When Adding Object References",
"x_gcve": [
{
"extensions": {
"bcp-05-x-01": {
"ai_annotations": [
{
"ai_level": "generated",
"description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
"gna_source": 1,
"models": [
{
"gna_source": 1,
"identifier": "qwen3.8:27b",
"name": "qwen3.8:27b",
"source": "ollama"
}
],
"review_status": "review",
"scope": "record",
"tags": [
"ai-computer-assisted:llm-generated",
"ai-computer-assisted:classification"
]
}
]
},
"bcp-05-x-02": {
"x_patch2vuln": {
"assumptions": [
"The exact affected and fixed version numbers are not specified in the patch metadata. The tag_version_boundary indicates the fix commit is 140 commits after v2.5.47, but the precise release version containing the fix is unknown.",
"The CAPEC-109 (Parameter Tampering) mapping is the closest available match but is not a perfect fit; the vulnerability is more precisely a missing granular authorization check rather than parameter manipulation. No CAPEC specifically models broken sub-entity access control.",
"The CVSS SC:H rating assumes that the exposed attribute/object data constitutes sensitive information (e.g., IOCs, malware indicators, threat intelligence) whose unauthorized disclosure is a significant confidentiality impact. If the data is considered low-sensitivity, SC:L may be more appropriate.",
"The vulnerability requires the attacker to already have event-level access; it does not grant access to events the user cannot see. The bypass is limited to granular restrictions within an already-accessible event.",
"The patch only addresses the add() method in ObjectReferencesController; other methods or controllers with similar patterns may or may not be affected, but no evidence of additional affected code paths is present in this patch."
],
"capecRationale": [
{
"capecId": "CAPEC-109",
"rationale": "The closest plausible CAPEC is Parameter Tampering. An authenticated user supplies an objectId parameter to the add() endpoint, and the application fails to enforce proper authorization on the sub-entities (attributes, objects, object attributes) associated with that object. The user effectively \u0027tampers\u0027 with the scope of accessible data by leveraging event-level access to reach restricted sub-entities. This is not a perfect match because the user is not modifying a parameter to an invalid value but rather exploiting the absence of a check on a valid parameter\u0027s sub-resources. No CAPEC specifically models \u0027broken granular access control\u0027 or \u0027insecure direct object reference at sub-entity level,\u0027 making CAPEC-109 the best available approximation."
}
],
"commit": "f36634e57b93ad9daa3ab9530b76b9e5cba796e2",
"confidence": "medium",
"credits": [
{
"lang": "en",
"type": "reporter",
"value": "Jeroen Pinoy"
},
{
"lang": "en",
"type": "remediation developer",
"value": "iglocska"
}
],
"cvssRationale": "AV:N: MISP is a network-accessible web application. AC:L: The attack requires no special conditions beyond having a valid account with event-level access; no race conditions or complex setup needed. AT:N: No manipulation of a separate attack target is required. PR:L: The attacker needs a low-privilege authenticated account (any non-site-admin user with access to an event containing restricted attributes/objects). UI:N: No victim interaction is required; the attacker simply calls the API endpoint. VC:N/VI:N/VA:N: The MISP server itself (the vulnerable component) is not compromised in its own confidentiality, integrity, or availability. SC:H: The resource impact is high because restricted attribute values, object metadata, and object-attribute data that are explicitly gated by distribution levels and sharing groups are exposed to an unauthorized user. SI:N: No integrity impact on the data. SA:N: No safety impact.",
"fixSummary": "The fix introduces granular authorization checks in the ObjectReferencesController add() method. For non-site-admin users, the query conditions for attributes, objects, and object attributes are now augmented with OR clauses that restrict results to: (1) entities belonging to the user\u0027s own event, (2) entities with unrestricted distribution levels (1, 2, 3, 5), or (3) entities with distribution level 4 whose sharing_group_id is in the user\u0027s authorized sharing group list. This ensures that only data the user is explicitly authorized to see under MISP\u0027s distribution and sharing-group model is included in the object reference operation.",
"generatedAt": "2026-09-21T13:15:28.001598Z",
"generator": "patch2vuln.py",
"model": "qwen3.8:27b",
"modelComparison": {
"rankings": [
{
"agreementScore": 9,
"assumptionCount": 5,
"confidence": "medium",
"model": "qwen3.8:27b",
"score": 5
}
],
"selectedModel": "qwen3.8:27b",
"selectionMethod": "deterministic-consensus-v1",
"selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
},
"patchSha256": "2688fca1334e3016b8b885ce926749f8d47e49d7b657cf38f7b2edcd6eaf6cf0",
"patchSummary": "In app/Controller/ObjectReferencesController.php, the add() method is modified to build dynamic condition arrays ($attributeConditions, $objectConditions, $objectAttributeConditions) that incorporate the current user\u0027s role and authorized sharing group IDs. For non-site-admin users, each condition array gains an OR clause filtering by event ownership, distribution levels (1,2,3,5), or distribution level 4 with a matching sharing_group_id from the user\u0027s authorized IDs. These condition arrays replace the previously hardcoded minimal conditions (deleted=0, object_id=0) in the Contain clauses for Attribute, Object, and nested Object.Attribute queries. The user\u0027s authorized sharing group IDs are obtained via SharingGroup-\u003eauthorizedIds($user).",
"patchTruncated": false,
"patches": [
{
"commit": "f36634e57b93ad9daa3ab9530b76b9e5cba796e2",
"patchSha256": "2688fca1334e3016b8b885ce926749f8d47e49d7b657cf38f7b2edcd6eaf6cf0",
"source": "https://github.com/MISP/MISP/commit/f36634e57.patch",
"sourceUrl": "https://github.com/MISP/MISP/commit/f36634e57.patch",
"subject": "fix: [security] Extended event granular restrictions ignored"
}
],
"source": "https://github.com/MISP/MISP/commit/f36634e57.patch",
"subject": "fix: [security] Extended event granular restrictions ignored",
"tagVersionBoundary": {
"commits_after_fix": 140,
"repository": "https://github.com/MISP/MISP",
"tag": "v2.5.47",
"version": "2.5.47",
"version_type": "semver"
},
"weaknessRationale": [
{
"cweId": "CWE-862",
"rationale": "The ObjectReferencesController add() method was missing authorization checks for granular distribution-level and sharing-group restrictions on attributes, objects, and object attributes. Only event-level access was verified, while the sub-entity level access controls (distribution levels 1-5, sharing group membership) were entirely absent from the query conditions. This is a classic missing authorization check at a finer granularity than what was enforced."
}
]
}
}
},
"recordType": "advisory",
"vulnId": "GCVE-1-2026-20106"
}
]
}
},
"cveMetadata": {
"assignerOrgId": "00000000-0000-4000-9000-000000000000",
"cveId": "CVE-2026-94394",
"datePublished": "2026-09-21T13:25:31.534037Z",
"dateReserved": "2026-09-21T13:25:38.483Z",
"dateUpdated": "2026-09-21T13:25:38.557803Z",
"state": "PUBLISHED",
"vulnId": "GCVE-1-2026-20106"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.1"
}
GCVE-1-2026-20057 (CVE-2026-94393)
Vulnerability from gna-1 – Published: 2026-09-21 13:13 – Updated: 2026-09-21 13:13| URL | Tags |
|---|---|
| https://github.com/MISP/MISP/commit/43665b9bb | patch |
qwen3.8:27b
advisory
bcp-05-x-01bcp-05-x-02
Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.
| Model | Source | Identifier |
|---|---|---|
| qwen3.8:27b | ollama | qwen3.8:27b |
- Generator
-
patch2vuln.pyon 2026-09-21 13:01 - Model
qwen3.8:27b- Input
-
https://github.com/MISP/MISP/commit/43665b9bb.patch
2306bd95d16b… - Confidence
- medium
| Commit | Subject | Patch SHA-256 |
|---|---|---|
43665b9bb6bd
|
fix: [security] Refuse to adopt an event report that belongs | 2306bd95d16b… |
Fix summary
The fix adds an ownership check in editReport: before adopting an existing report found by UUID, the code now verifies that the report's event_id matches the event being edited. If the UUID resolves to a report belonging to a different event, the operation is rejected with an error message, preventing cross-event reparenting, unauthorized read, and unauthorized overwrite of reports.
Patch summary
In app/Model/EventReport.php, the editReport method's else-branch (handling an existing report found by UUID) now includes an 11-line guard: it compares (string)$existingReport['EventReport']['event_id'] against (string)$eventId and, on mismatch, appends an error string and returns early, preventing the subsequent assignment of the foreign report's ID to the current event's report record.
CVSS rationale
AV:N: MISP is a network-accessible web application. AC:L: the attacker only needs to supply a known/guessed UUID in a standard edit request; no race or complex condition. AT:N: no manipulation of the target environment is required. PR:L: the attacker must be an authenticated user with editor role on at least one event. UI:N: no victim interaction is needed. VC:N/VI:N/VA:N: the MISP application itself (its configuration, credentials, availability) is not directly impacted. SC:H: the attacker gains read access to private report data in events they cannot otherwise see. SI:H: the attacker can overwrite and modify report data belonging to other events. SA:N: no safety-system impact.
Weakness rationale
- CWE-639 The report UUID is a user-controlled key resolved globally without scoping to the caller's event context. The system failed to verify that the referenced object (report) belongs to the event the caller is authorized to edit, allowing cross-event access.
- CWE-284 The broader category applies: MISP's per-event access control model was bypassed because the editReport path did not enforce that a report UUID maps to the same event the caller is operating on.
Attack pattern rationale
- CAPEC-174 The edge case of a UUID belonging to a different event was not handled: the code assumed any UUID found in the global table belonged to the current event. The attacker exploits this unhandled edge case to reparent a foreign report. This is the closest CAPEC; a more specific 'IDOR' or 'broken object reference' CAPEC does not exist in the CAPEC catalog, so CAPEC-174 is the best available match.
Assumptions to verify
- The affected version range is inferred from the tag_version_boundary (v2.5.47, 56 commits after fix); the exact first affected and first fixed release versions are not stated in the patch and are marked unspecified.
- PR:L assumes the attacker needs at minimum editor-level access to one event; the patch does not specify whether viewer-level access could also trigger the code path, but the commit message explicitly references 'an event editor'.
- CAPEC-174 is selected as the closest available attack pattern; no CAPEC entry specifically covers IDOR or broken object reference resolution, so the mapping is approximate.
- The CVSS sub-component impact (SC:H, SI:H) assumes the event report data constitutes a distinct data asset whose confidentiality and integrity are the primary impact, rather than the MISP application's own state.
- The commit date (2026-09-16) and tag boundary are taken at face value from the supplied metadata; no independent verification of release timing was performed.
Model comparison
Selected qwen3.8:27b
by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.
| Model | Score | Agreement | Confidence | Assumptions |
|---|---|---|---|---|
qwen3.8:27b |
5 | 9 | medium | 5 |
{
"containers": {
"cna": {
"affected": [
{
"modules": [
"app/Model/EventReport.php"
],
"product": "MISP",
"programFiles": [
"app/Model/EventReport.php"
],
"repo": "https://github.com/MISP/MISP",
"vendor": "MISP",
"versions": [
{
"lessThan": "2.5.47",
"status": "affected",
"version": "0",
"versionType": "semver"
}
]
}
],
"credits": [
{
"lang": "en",
"type": "remediation developer",
"value": "iglocska"
},
{
"lang": "en",
"type": "remediation developer",
"value": "Claude Opus 4.8"
},
{
"lang": "en",
"type": "reporter",
"value": "David Andr\u00e9"
},
{
"lang": "en",
"type": "reporter",
"value": "Jeroen Pinoy"
}
],
"descriptions": [
{
"lang": "en",
"supportingMedia": [
{
"base64": false,
"type": "text/html",
"value": "\u003cdiv\u003e\u003c/div\u003e\u003cp\u003eWhen a user creates or edits a report inside an event, MISP can identify an existing report using its UUID without properly checking whether that report actually belongs to the same event.\u003c/p\u003e\u003cp\u003eAs a result, a user who has editing rights on one event could potentially move a report from another event into their own event, as long as they know or can guess the report\u2019s UUID. Once moved, they could view and change information that they were not originally allowed to access.\u003c/p\u003e\u003cp\u003eThe vulnerability requires the attacker to have editor access to at least one event and to know or discover a valid report UUID.\u003c/p\u003e\u003cp\u003eThe main impact is that private event reports could be exposed or modified across event boundaries, bypassing MISP\u2019s normal access restrictions.\u003c/p\u003e\u003cp\u003eVersion affected: \u0026lt;2.5.47\u003c/p\u003e\u003cdiv\u003e\u003c/div\u003e"
}
],
"value": "When a user creates or edits a report inside an event, MISP can identify an existing report using its UUID without properly checking whether that report actually belongs to the same event.\n\nAs a result, a user who has editing rights on one event could potentially move a report from another event into their own event, as long as they know or can guess the report\u2019s UUID. Once moved, they could view and change information that they were not originally allowed to access.\n\nThe vulnerability requires the attacker to have editor access to at least one event and to know or discover a valid report UUID.\n\nThe main impact is that private event reports could be exposed or modified across event boundaries, bypassing MISP\u2019s normal access restrictions.\n\nVersion affected: \u003c2.5.47"
}
],
"impacts": [
{
"capecId": "CAPEC-174",
"descriptions": [
{
"lang": "en",
"value": "CAPEC-174 Exploiting Incorrectly Handled Edge Cases"
}
]
}
],
"metrics": [
{
"cvssV4_0": {
"Automatable": "NOT_DEFINED",
"Recovery": "NOT_DEFINED",
"Safety": "NOT_DEFINED",
"attackComplexity": "LOW",
"attackRequirements": "NONE",
"attackVector": "NETWORK",
"baseScore": 6.4,
"baseSeverity": "MEDIUM",
"privilegesRequired": "LOW",
"providerUrgency": "NOT_DEFINED",
"subAvailabilityImpact": "NONE",
"subConfidentialityImpact": "HIGH",
"subIntegrityImpact": "HIGH",
"userInteraction": "NONE",
"valueDensity": "NOT_DEFINED",
"vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:N/VI:N/VA:N/SC:H/SI:H/SA:N",
"version": "4.0",
"vulnAvailabilityImpact": "NONE",
"vulnConfidentialityImpact": "NONE",
"vulnIntegrityImpact": "NONE",
"vulnerabilityResponseEffort": "NOT_DEFINED"
},
"format": "CVSS",
"scenarios": [
{
"lang": "en",
"value": "GENERAL"
}
]
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-639",
"description": "CWE-639 Authorization Bypass Through User-Controlled Key",
"lang": "en",
"type": "CWE"
}
]
},
{
"descriptions": [
{
"cweId": "CWE-284",
"description": "CWE-284 Improper Access Control",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"orgId": "00000000-0000-4000-9000-000000000000"
},
"references": [
{
"name": "Security patch",
"tags": [
"patch"
],
"url": "https://github.com/MISP/MISP/commit/43665b9bb"
}
],
"solutions": [
{
"lang": "en",
"supportingMedia": [
{
"base64": false,
"type": "text/html",
"value": "\u003cp\u003eThe fix adds an ownership check in editReport: before adopting an existing report found by UUID, the code now verifies that the report\u0027s event_id matches the event being edited. If the UUID resolves to a report belonging to a different event, the operation is rejected with an error message, preventing cross-event reparenting, unauthorized read, and unauthorized overwrite of reports.\u003c/p\u003e"
}
],
"value": "The fix adds an ownership check in editReport: before adopting an existing report found by UUID, the code now verifies that the report\u0027s event_id matches the event being edited. If the UUID resolves to a report belonging to a different event, the operation is rejected with an error message, preventing cross-event reparenting, unauthorized read, and unauthorized overwrite of reports."
}
],
"title": "MISP Event Report Cross-Event Reparenting via Unscoped UUID Resolution in editReport",
"x_gcve": [
{
"extensions": {
"bcp-05-x-01": {
"ai_annotations": [
{
"ai_level": "generated",
"description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
"gna_source": 1,
"models": [
{
"gna_source": 1,
"identifier": "qwen3.8:27b",
"name": "qwen3.8:27b",
"source": "ollama"
}
],
"review_status": "review",
"scope": "record",
"tags": [
"ai-computer-assisted:llm-generated",
"ai-computer-assisted:classification"
]
}
]
},
"bcp-05-x-02": {
"x_patch2vuln": {
"assumptions": [
"The affected version range is inferred from the tag_version_boundary (v2.5.47, 56 commits after fix); the exact first affected and first fixed release versions are not stated in the patch and are marked unspecified.",
"PR:L assumes the attacker needs at minimum editor-level access to one event; the patch does not specify whether viewer-level access could also trigger the code path, but the commit message explicitly references \u0027an event editor\u0027.",
"CAPEC-174 is selected as the closest available attack pattern; no CAPEC entry specifically covers IDOR or broken object reference resolution, so the mapping is approximate.",
"The CVSS sub-component impact (SC:H, SI:H) assumes the event report data constitutes a distinct data asset whose confidentiality and integrity are the primary impact, rather than the MISP application\u0027s own state.",
"The commit date (2026-09-16) and tag boundary are taken at face value from the supplied metadata; no independent verification of release timing was performed."
],
"capecRationale": [
{
"capecId": "CAPEC-174",
"rationale": "The edge case of a UUID belonging to a different event was not handled: the code assumed any UUID found in the global table belonged to the current event. The attacker exploits this unhandled edge case to reparent a foreign report. This is the closest CAPEC; a more specific \u0027IDOR\u0027 or \u0027broken object reference\u0027 CAPEC does not exist in the CAPEC catalog, so CAPEC-174 is the best available match."
}
],
"commit": "43665b9bb6bd39af0db1f38e608bdf2cb84b9dec",
"confidence": "medium",
"credits": [
{
"lang": "en",
"type": "remediation developer",
"value": "iglocska"
},
{
"lang": "en",
"type": "remediation developer",
"value": "Claude Opus 4.8"
},
{
"lang": "en",
"type": "reporter",
"value": "David Andr\u00e9"
},
{
"lang": "en",
"type": "reporter",
"value": "Jeroen Pinoy"
}
],
"cvssRationale": "AV:N: MISP is a network-accessible web application. AC:L: the attacker only needs to supply a known/guessed UUID in a standard edit request; no race or complex condition. AT:N: no manipulation of the target environment is required. PR:L: the attacker must be an authenticated user with editor role on at least one event. UI:N: no victim interaction is needed. VC:N/VI:N/VA:N: the MISP application itself (its configuration, credentials, availability) is not directly impacted. SC:H: the attacker gains read access to private report data in events they cannot otherwise see. SI:H: the attacker can overwrite and modify report data belonging to other events. SA:N: no safety-system impact.",
"fixSummary": "The fix adds an ownership check in editReport: before adopting an existing report found by UUID, the code now verifies that the report\u0027s event_id matches the event being edited. If the UUID resolves to a report belonging to a different event, the operation is rejected with an error message, preventing cross-event reparenting, unauthorized read, and unauthorized overwrite of reports.",
"generatedAt": "2026-09-21T13:01:16.364280Z",
"generator": "patch2vuln.py",
"model": "qwen3.8:27b",
"modelComparison": {
"rankings": [
{
"agreementScore": 9,
"assumptionCount": 5,
"confidence": "medium",
"model": "qwen3.8:27b",
"score": 5
}
],
"selectedModel": "qwen3.8:27b",
"selectionMethod": "deterministic-consensus-v1",
"selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
},
"patchSha256": "2306bd95d16b9c00582997897aa9181253364082a09589640fce576a03beafb4",
"patchSummary": "In app/Model/EventReport.php, the editReport method\u0027s else-branch (handling an existing report found by UUID) now includes an 11-line guard: it compares (string)$existingReport[\u0027EventReport\u0027][\u0027event_id\u0027] against (string)$eventId and, on mismatch, appends an error string and returns early, preventing the subsequent assignment of the foreign report\u0027s ID to the current event\u0027s report record.",
"patchTruncated": false,
"patches": [
{
"commit": "43665b9bb6bd39af0db1f38e608bdf2cb84b9dec",
"patchSha256": "2306bd95d16b9c00582997897aa9181253364082a09589640fce576a03beafb4",
"source": "https://github.com/MISP/MISP/commit/43665b9bb.patch",
"sourceUrl": "https://github.com/MISP/MISP/commit/43665b9bb.patch",
"subject": "fix: [security] Refuse to adopt an event report that belongs"
}
],
"source": "https://github.com/MISP/MISP/commit/43665b9bb.patch",
"subject": "fix: [security] Refuse to adopt an event report that belongs",
"tagVersionBoundary": {
"commits_after_fix": 56,
"repository": "https://github.com/MISP/MISP",
"tag": "v2.5.47",
"version": "2.5.47",
"version_type": "semver"
},
"weaknessRationale": [
{
"cweId": "CWE-639",
"rationale": "The report UUID is a user-controlled key resolved globally without scoping to the caller\u0027s event context. The system failed to verify that the referenced object (report) belongs to the event the caller is authorized to edit, allowing cross-event access."
},
{
"cweId": "CWE-284",
"rationale": "The broader category applies: MISP\u0027s per-event access control model was bypassed because the editReport path did not enforce that a report UUID maps to the same event the caller is operating on."
}
]
}
}
},
"recordType": "advisory",
"vulnId": "GCVE-1-2026-20057"
}
]
}
},
"cveMetadata": {
"assignerOrgId": "00000000-0000-4000-9000-000000000000",
"cveId": "CVE-2026-94393",
"datePublished": "2026-09-21T13:13:46.709115Z",
"dateReserved": "2026-09-21T13:13:54.299Z",
"dateUpdated": "2026-09-21T13:13:54.565172Z",
"state": "PUBLISHED",
"vulnId": "GCVE-1-2026-20057"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.1"
}
GCVE-1-2026-20009 (CVE-2026-94383)
Vulnerability from gna-1 – Published: 2026-09-21 12:51 – Updated: 2026-09-21 12:52| URL | Tags |
|---|---|
| https://github.com/MISP/MISP/commit/e5a943178 | patch |
qwen3.8:27b
advisory
bcp-05-x-01bcp-05-x-02
Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.
| Model | Source | Identifier |
|---|---|---|
| qwen3.8:27b | ollama | qwen3.8:27b |
- Generator
-
patch2vuln.pyon 2026-09-21 12:44 - Model
qwen3.8:27b- Input
-
https://github.com/MISP/MISP/commit/e5a943178.patch
9b7cc757cbbc… - Confidence
- medium
| Commit | Subject | Patch SHA-256 |
|---|---|---|
e5a9431784b4
|
fix: [security] arbitrary script execution via insufficient | 9b7cc757cbbc… |
Fix summary
The vulnerability is remediated by enforcing a strict whitelist of allowed file extensions (txt, list, blocklist, csv) on the blocklist filename parameter before the file is processed. Any filename with an extension outside this whitelist is rejected with an error, preventing an attacker from placing files with executable or dangerous extensions (e.g., .php, .phtml) into the MISP export directory where they could be interpreted by the web server.
Patch summary
In app/Model/WorkflowModules/action/Module_blocklist_action.php, five lines are inserted after the existing basename() call. The added code defines an $allowedExtensions array containing 'txt', 'list', 'blocklist', and 'csv', extracts the file extension via pathinfo() with PATHINFO_EXTENSION, and performs a strict (===) in_array() check. If the extension is not in the whitelist, the method returns early with a failure message 'Blocklist filename has an invalid extension.' This check is placed before the existing empty/dot filename validation.
CVSS rationale
AV:N: MISP is a network-accessible web application. AC:L: The attack requires only specifying a filename with a dangerous extension in a workflow parameter; no race conditions or complex timing are needed. AT:N: No manipulation of the target system is required beyond the normal API call. PR:H: The commit message explicitly states 'site admin only', requiring high-privilege MISP access. UI:N: No victim interaction is needed; the workflow action is triggered by the admin themselves. VC:H, VI:H, VA:H: Arbitrary script execution on the MISP server compromises all three local security properties. SC:N, SI:N, SA:N: The impact is confined to the MISP server; no evidence of lateral impact on other systems is present in the patch.
Weakness rationale
- CWE-434 The blocklist filename parameter accepted arbitrary file extensions without restriction. A site admin could supply a filename with a dangerous extension (e.g., .php) that would be placed in a web-accessible export directory, enabling script execution. The fix restricts extensions to a safe whitelist, which is the canonical remediation for CWE-434.
- CWE-20 The filename parameter was not validated against a set of acceptable values (extensions). Only path stripping and empty-string checks were performed, leaving the extension entirely unvalidated. This is a general improper input validation weakness that enabled the more specific dangerous-file-type issue.
Attack pattern rationale
- CAPEC-170 CAPEC-170 describes an attacker uploading or placing a file with a dangerous extension into a target directory where it can be executed. In this case, a site admin specifies a blocklist filename with a dangerous extension (e.g., .php) that is placed in the MISP export directory. The mechanism is not a traditional multipart upload but rather a workflow parameter that results in a file being written to a web-accessible location. CAPEC-170 is the closest available pattern; the uncertainty is that the file is not uploaded via a standard upload endpoint but is instead referenced by name in a workflow action, making the mapping approximate.
Assumptions to verify
- The commit message states 'site admin only', which is taken as evidence that the vulnerable workflow module requires MISP site-administrator privileges. The exact privilege model and whether lower-privilege roles could reach this code path is not confirmed by the patch alone.
- The assumption that the MISP export directory is web-accessible and that the web server is configured to execute scripts (e.g., PHP) from that directory is inferred from the commit subject 'arbitrary script execution'. The patch does not include web server configuration evidence.
- The affected version boundary (prior to v2.5.47) is inferred from the tag_version_boundary metadata showing v2.5.47 as the nearest tag with 142 commits after the fix. The exact first affected version is not stated in the patch.
- CAPEC-170 (File Upload) is the closest available attack pattern, but the actual mechanism is a workflow parameter specifying a filename rather than a traditional file upload. This mapping is approximate.
- The CVSS PR:H rating assumes that 'site admin' in MISP corresponds to a high-privilege role requiring authentication and elevated permissions, consistent with MISP's role model.
Model comparison
Selected qwen3.8:27b
by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.
| Model | Score | Agreement | Confidence | Assumptions |
|---|---|---|---|---|
qwen3.8:27b |
5 | 9 | medium | 5 |
{
"containers": {
"cna": {
"affected": [
{
"modules": [
"WorkflowModules/action/Module_blocklist_action"
],
"product": "MISP",
"programFiles": [
"app/Model/WorkflowModules/action/Module_blocklist_action.php"
],
"repo": "https://github.com/MISP/MISP",
"vendor": "MISP",
"versions": [
{
"lessThan": "2.5.47",
"status": "affected",
"version": "0",
"versionType": "semver"
}
]
}
],
"credits": [
{
"lang": "en",
"type": "reporter",
"value": "Jeroen Pinoy"
},
{
"lang": "en",
"type": "remediation developer",
"value": "iglocska"
}
],
"descriptions": [
{
"lang": "en",
"supportingMedia": [
{
"base64": false,
"type": "text/html",
"value": "\u003cp\u003eThe MISP blocklist workflow module accepted a user-supplied blocklist filename parameter without validating the file extension. The only sanitization applied was basename() to strip path components and a check for empty or dot values. A site administrator could specify a filename with an arbitrary extension that would be placed in the MISP export directory. If the underlying web server is configured to interpret and execute scripts from that directory, the resulting file could be invoked, leading to arbitrary code execution in the context of the web server process.\u003c/p\u003e\u003cp\u003eThe vulnerability requires the attacker to hold site-administrator privileges within MISP, as the blocklist workflow module is restricted to that role. No additional user interaction is required beyond triggering the workflow action with a crafted filename parameter. The impact is full compromise of the MISP server\u0027s confidentiality, integrity, and availability, as arbitrary script execution grants the attacker the same privileges as the web server user.\u003c/p\u003e\u003cp\u003eVersion affected: \u0026lt;2.5.47\u003c/p\u003e"
}
],
"value": "The MISP blocklist workflow module accepted a user-supplied blocklist filename parameter without validating the file extension. The only sanitization applied was basename() to strip path components and a check for empty or dot values. A site administrator could specify a filename with an arbitrary extension that would be placed in the MISP export directory. If the underlying web server is configured to interpret and execute scripts from that directory, the resulting file could be invoked, leading to arbitrary code execution in the context of the web server process.\n\nThe vulnerability requires the attacker to hold site-administrator privileges within MISP, as the blocklist workflow module is restricted to that role. No additional user interaction is required beyond triggering the workflow action with a crafted filename parameter. The impact is full compromise of the MISP server\u0027s confidentiality, integrity, and availability, as arbitrary script execution grants the attacker the same privileges as the web server user.\n\nVersion affected: \u003c2.5.47"
}
],
"impacts": [
{
"capecId": "CAPEC-170",
"descriptions": [
{
"lang": "en",
"value": "CAPEC-170 File Upload"
}
]
}
],
"metrics": [
{
"cvssV4_0": {
"Automatable": "NOT_DEFINED",
"Recovery": "NOT_DEFINED",
"Safety": "NOT_DEFINED",
"attackComplexity": "LOW",
"attackRequirements": "NONE",
"attackVector": "NETWORK",
"baseScore": 8.6,
"baseSeverity": "HIGH",
"privilegesRequired": "HIGH",
"providerUrgency": "NOT_DEFINED",
"subAvailabilityImpact": "NONE",
"subConfidentialityImpact": "NONE",
"subIntegrityImpact": "NONE",
"userInteraction": "NONE",
"valueDensity": "NOT_DEFINED",
"vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:H/UI:N/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N",
"version": "4.0",
"vulnAvailabilityImpact": "HIGH",
"vulnConfidentialityImpact": "HIGH",
"vulnIntegrityImpact": "HIGH",
"vulnerabilityResponseEffort": "NOT_DEFINED"
},
"format": "CVSS",
"scenarios": [
{
"lang": "en",
"value": "GENERAL"
}
]
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-434",
"description": "CWE-434 Unrestricted Upload of File with Dangerous Type",
"lang": "en",
"type": "CWE"
}
]
},
{
"descriptions": [
{
"cweId": "CWE-20",
"description": "CWE-20 Improper Input Validation",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"orgId": "00000000-0000-4000-9000-000000000000"
},
"references": [
{
"name": "Security patch",
"tags": [
"patch"
],
"url": "https://github.com/MISP/MISP/commit/e5a943178"
}
],
"solutions": [
{
"lang": "en",
"supportingMedia": [
{
"base64": false,
"type": "text/html",
"value": "\u003cp\u003eThe vulnerability is remediated by enforcing a strict whitelist of allowed file extensions (txt, list, blocklist, csv) on the blocklist filename parameter before the file is processed. Any filename with an extension outside this whitelist is rejected with an error, preventing an attacker from placing files with executable or dangerous extensions (e.g., .php, .phtml) into the MISP export directory where they could be interpreted by the web server.\u003c/p\u003e"
}
],
"value": "The vulnerability is remediated by enforcing a strict whitelist of allowed file extensions (txt, list, blocklist, csv) on the blocklist filename parameter before the file is processed. Any filename with an extension outside this whitelist is rejected with an error, preventing an attacker from placing files with executable or dangerous extensions (e.g., .php, .phtml) into the MISP export directory where they could be interpreted by the web server."
}
],
"title": "MISP Blocklist Workflow Module: Arbitrary Script Execution via Unrestricted File Extension",
"x_gcve": [
{
"extensions": {
"bcp-05-x-01": {
"ai_annotations": [
{
"ai_level": "generated",
"description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
"gna_source": 1,
"models": [
{
"gna_source": 1,
"identifier": "qwen3.8:27b",
"name": "qwen3.8:27b",
"source": "ollama"
}
],
"review_status": "review",
"scope": "record",
"tags": [
"ai-computer-assisted:llm-generated",
"ai-computer-assisted:classification"
]
}
]
},
"bcp-05-x-02": {
"x_patch2vuln": {
"assumptions": [
"The commit message states \u0027site admin only\u0027, which is taken as evidence that the vulnerable workflow module requires MISP site-administrator privileges. The exact privilege model and whether lower-privilege roles could reach this code path is not confirmed by the patch alone.",
"The assumption that the MISP export directory is web-accessible and that the web server is configured to execute scripts (e.g., PHP) from that directory is inferred from the commit subject \u0027arbitrary script execution\u0027. The patch does not include web server configuration evidence.",
"The affected version boundary (prior to v2.5.47) is inferred from the tag_version_boundary metadata showing v2.5.47 as the nearest tag with 142 commits after the fix. The exact first affected version is not stated in the patch.",
"CAPEC-170 (File Upload) is the closest available attack pattern, but the actual mechanism is a workflow parameter specifying a filename rather than a traditional file upload. This mapping is approximate.",
"The CVSS PR:H rating assumes that \u0027site admin\u0027 in MISP corresponds to a high-privilege role requiring authentication and elevated permissions, consistent with MISP\u0027s role model."
],
"capecRationale": [
{
"capecId": "CAPEC-170",
"rationale": "CAPEC-170 describes an attacker uploading or placing a file with a dangerous extension into a target directory where it can be executed. In this case, a site admin specifies a blocklist filename with a dangerous extension (e.g., .php) that is placed in the MISP export directory. The mechanism is not a traditional multipart upload but rather a workflow parameter that results in a file being written to a web-accessible location. CAPEC-170 is the closest available pattern; the uncertainty is that the file is not uploaded via a standard upload endpoint but is instead referenced by name in a workflow action, making the mapping approximate."
}
],
"commit": "e5a9431784b4a78faddc0cc6efc2a287f2fd3d08",
"confidence": "medium",
"credits": [
{
"lang": "en",
"type": "reporter",
"value": "Jeroen Pinoy"
},
{
"lang": "en",
"type": "remediation developer",
"value": "iglocska"
}
],
"cvssRationale": "AV:N: MISP is a network-accessible web application. AC:L: The attack requires only specifying a filename with a dangerous extension in a workflow parameter; no race conditions or complex timing are needed. AT:N: No manipulation of the target system is required beyond the normal API call. PR:H: The commit message explicitly states \u0027site admin only\u0027, requiring high-privilege MISP access. UI:N: No victim interaction is needed; the workflow action is triggered by the admin themselves. VC:H, VI:H, VA:H: Arbitrary script execution on the MISP server compromises all three local security properties. SC:N, SI:N, SA:N: The impact is confined to the MISP server; no evidence of lateral impact on other systems is present in the patch.",
"fixSummary": "The vulnerability is remediated by enforcing a strict whitelist of allowed file extensions (txt, list, blocklist, csv) on the blocklist filename parameter before the file is processed. Any filename with an extension outside this whitelist is rejected with an error, preventing an attacker from placing files with executable or dangerous extensions (e.g., .php, .phtml) into the MISP export directory where they could be interpreted by the web server.",
"generatedAt": "2026-09-21T12:44:28.234229Z",
"generator": "patch2vuln.py",
"model": "qwen3.8:27b",
"modelComparison": {
"rankings": [
{
"agreementScore": 9,
"assumptionCount": 5,
"confidence": "medium",
"model": "qwen3.8:27b",
"score": 5
}
],
"selectedModel": "qwen3.8:27b",
"selectionMethod": "deterministic-consensus-v1",
"selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
},
"patchSha256": "9b7cc757cbbc8a4dd7491c6388e0ee12057c790bb6ca21e10b1d78209bbb7748",
"patchSummary": "In app/Model/WorkflowModules/action/Module_blocklist_action.php, five lines are inserted after the existing basename() call. The added code defines an $allowedExtensions array containing \u0027txt\u0027, \u0027list\u0027, \u0027blocklist\u0027, and \u0027csv\u0027, extracts the file extension via pathinfo() with PATHINFO_EXTENSION, and performs a strict (===) in_array() check. If the extension is not in the whitelist, the method returns early with a failure message \u0027Blocklist filename has an invalid extension.\u0027 This check is placed before the existing empty/dot filename validation.",
"patchTruncated": false,
"patches": [
{
"commit": "e5a9431784b4a78faddc0cc6efc2a287f2fd3d08",
"patchSha256": "9b7cc757cbbc8a4dd7491c6388e0ee12057c790bb6ca21e10b1d78209bbb7748",
"source": "https://github.com/MISP/MISP/commit/e5a943178.patch",
"sourceUrl": "https://github.com/MISP/MISP/commit/e5a943178.patch",
"subject": "fix: [security] arbitrary script execution via insufficient"
}
],
"source": "https://github.com/MISP/MISP/commit/e5a943178.patch",
"subject": "fix: [security] arbitrary script execution via insufficient",
"tagVersionBoundary": {
"commits_after_fix": 142,
"repository": "https://github.com/MISP/MISP",
"tag": "v2.5.47",
"version": "2.5.47",
"version_type": "semver"
},
"weaknessRationale": [
{
"cweId": "CWE-434",
"rationale": "The blocklist filename parameter accepted arbitrary file extensions without restriction. A site admin could supply a filename with a dangerous extension (e.g., .php) that would be placed in a web-accessible export directory, enabling script execution. The fix restricts extensions to a safe whitelist, which is the canonical remediation for CWE-434."
},
{
"cweId": "CWE-20",
"rationale": "The filename parameter was not validated against a set of acceptable values (extensions). Only path stripping and empty-string checks were performed, leaving the extension entirely unvalidated. This is a general improper input validation weakness that enabled the more specific dangerous-file-type issue."
}
]
}
}
},
"recordType": "advisory",
"vulnId": "GCVE-1-2026-20009"
}
]
}
},
"cveMetadata": {
"assignerOrgId": "00000000-0000-4000-9000-000000000000",
"cveId": "CVE-2026-94383",
"datePublished": "2026-09-21T12:51:51.619869Z",
"dateReserved": "2026-09-21T12:52:01.101Z",
"dateUpdated": "2026-09-21T12:52:01.169089Z",
"state": "PUBLISHED",
"vulnId": "GCVE-1-2026-20009"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.1"
}
GCVE-1-2026-20048 (CVE-2026-94381)
Vulnerability from gna-1 – Published: 2026-09-21 12:42 – Updated: 2026-09-21 12:42- CWE-269 - Improper Privilege Management
| URL | Tags |
|---|---|
| https://github.com/MISP/MISP/commit/fd27e592a | patch |
qwen3.8:27b
advisory
bcp-05-x-01bcp-05-x-02
Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.
| Model | Source | Identifier |
|---|---|---|
| qwen3.8:27b | ollama | qwen3.8:27b |
- Generator
-
patch2vuln.pyon 2026-09-21 12:37 - Model
qwen3.8:27b- Input
-
https://github.com/MISP/MISP/commit/fd27e592a.patch
d4143841f5f0… - Confidence
- medium
| Commit | Subject | Patch SHA-256 |
|---|---|---|
fd27e592a617
|
fix: [security] read only api keys can regain full role | d4143841f5f0… |
Fix summary
The fix removes the call to User::getAuthUser() and Auth::login() from the updateLoginTime() method, which was re-authenticating the user with their full role permissions. It replaces this with a call to _refreshAuth(), which refreshes the authentication state while preserving the current (restricted) permission level associated with the API key, preventing privilege escalation.
Patch summary
In app/Controller/UsersController.php, the updateLoginTime() method is modified: the lines that fetched the user record via User::find(), called User::getAuthUser() to retrieve the full auth user, and called Auth::login() to re-establish the session are removed. These are replaced with a single call to $this->_refreshAuth(), which updates the authentication context without upgrading the permission scope. Net change: 1 insertion, 6 deletions.
CVSS rationale
AV:N - the endpoint is accessible over the network via HTTP POST. AC:L - the attack requires only a single POST request to a known endpoint; no race conditions or complex timing. AT:N - no manipulation of the target system is needed beyond the API call. PR:L - the attacker must possess a valid read-only API key (authenticated user). UI:N - no user interaction is required. VC:H - the escalated user can read all data in the MISP instance that their full role permits. VI:H - the escalated user can create, modify, and delete threat intelligence data. VA:H - the escalated user can delete data, modify configurations, or disrupt the MISP service. SC/SI/SA:N - the impact is confined to the MISP instance itself; no cascading impact on other systems is evidenced.
Weakness rationale
- CWE-269 The system fails to maintain the restricted privilege level (read-only) associated with the API key authentication. The updateLoginTime endpoint re-establishes the session with the user's full role permissions, effectively granting higher privileges than the authentication credential intended. This is a failure to properly manage and enforce the privilege boundary during session refresh.
Attack pattern rationale
- CAPEC-114 The vulnerability arises because the updateLoginTime endpoint incorrectly reconfigures the user's privilege level during session refresh, upgrading from the API-key-scoped read-only permissions to the user's full role. The attacker exploits this misconfiguration by simply calling the endpoint. CAPEC-114 is the closest match as it covers scenarios where an application incorrectly assigns or maintains privilege levels. Uncertainty: no CAPEC specifically covers 'session privilege escalation via auth refresh,' so this is the best available approximation.
Assumptions to verify
- The tag_version_boundary indicates the fix commit is 141 commits before v2.5.47, implying v2.5.47 contains the fix; versions prior to v2.5.47 are assumed affected. No explicit version range is stated in the patch.
- The _refreshAuth() method is assumed to preserve the current API-key-scoped permission level rather than re-fetching full role permissions; this is inferred from the commit subject and the removal of getAuthUser()/Auth::login().
- The 'full role' referenced in the commit subject is assumed to be a higher-privilege role (e.g., admin or site admin) compared to the read-only API key scope.
- CAPEC-114 is the closest available attack pattern; no CAPEC specifically models session privilege escalation via authentication refresh, so this mapping carries some uncertainty.
- CVSS impact metrics (VC/VI/VA) assume the user's full role grants broad administrative access to the MISP instance; actual impact depends on the specific role assigned to the user.
Model comparison
Selected qwen3.8:27b
by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.
| Model | Score | Agreement | Confidence | Assumptions |
|---|---|---|---|---|
qwen3.8:27b |
5 | 9 | medium | 5 |
{
"containers": {
"cna": {
"affected": [
{
"modules": [
"UsersController"
],
"product": "MISP",
"programFiles": [
"app/Controller/UsersController.php"
],
"repo": "https://github.com/MISP/MISP",
"vendor": "MISP",
"versions": [
{
"lessThan": "2.5.47",
"status": "affected",
"version": "0",
"versionType": "semver"
}
]
}
],
"credits": [
{
"lang": "en",
"type": "reporter",
"value": "Jeroen Pinoy"
},
{
"lang": "en",
"type": "remediation developer",
"value": "iglocska"
}
],
"descriptions": [
{
"lang": "en",
"supportingMedia": [
{
"base64": false,
"type": "text/html",
"value": "\u003cp\u003e\u003c/p\u003e\u003cp\u003eMISP has a security issue that can let a user gain more access than their API key is supposed to allow.\u003c/p\u003e\u003cp\u003eA read-only API key should only let someone view information. However, after logging in with such a key, a specific MISP function could accidentally restore the user\u2019s normal account permissions. This means someone with a read-only API key could potentially gain write, delete, or even administrator access if their underlying account has those permissions.\u003c/p\u003e\u003cp\u003eExploiting the issue requires a valid read-only API key and a single request to the affected function.\u003c/p\u003e\u003cp\u003eThe main impact is that MISP\u2019s API key restrictions can be bypassed, allowing actions that the API key was specifically meant to prevent.\u003c/p\u003e\u003cbr\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eVersion affected: \u0026lt;2.5.47\u003c/p\u003e"
}
],
"value": "MISP has a security issue that can let a user gain more access than their API key is supposed to allow.\n\nA read-only API key should only let someone view information. However, after logging in with such a key, a specific MISP function could accidentally restore the user\u2019s normal account permissions. This means someone with a read-only API key could potentially gain write, delete, or even administrator access if their underlying account has those permissions.\n\nExploiting the issue requires a valid read-only API key and a single request to the affected function.\n\nThe main impact is that MISP\u2019s API key restrictions can be bypassed, allowing actions that the API key was specifically meant to prevent.\n\n\n\n\nVersion affected: \u003c2.5.47"
}
],
"impacts": [
{
"capecId": "CAPEC-114",
"descriptions": [
{
"lang": "en",
"value": "CAPEC-114 Exploiting Incorrectly Configured Privileges"
}
]
}
],
"metrics": [
{
"cvssV4_0": {
"Automatable": "NOT_DEFINED",
"Recovery": "NOT_DEFINED",
"Safety": "NOT_DEFINED",
"attackComplexity": "LOW",
"attackRequirements": "NONE",
"attackVector": "NETWORK",
"baseScore": 8.7,
"baseSeverity": "HIGH",
"privilegesRequired": "LOW",
"providerUrgency": "NOT_DEFINED",
"subAvailabilityImpact": "NONE",
"subConfidentialityImpact": "NONE",
"subIntegrityImpact": "NONE",
"userInteraction": "NONE",
"valueDensity": "NOT_DEFINED",
"vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N",
"version": "4.0",
"vulnAvailabilityImpact": "HIGH",
"vulnConfidentialityImpact": "HIGH",
"vulnIntegrityImpact": "HIGH",
"vulnerabilityResponseEffort": "NOT_DEFINED"
},
"format": "CVSS",
"scenarios": [
{
"lang": "en",
"value": "GENERAL"
}
]
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-269",
"description": "CWE-269 Improper Privilege Management",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"orgId": "00000000-0000-4000-9000-000000000000"
},
"references": [
{
"name": "Security patch",
"tags": [
"patch"
],
"url": "https://github.com/MISP/MISP/commit/fd27e592a"
}
],
"solutions": [
{
"lang": "en",
"supportingMedia": [
{
"base64": false,
"type": "text/html",
"value": "\u003cp\u003eThe fix removes the call to User::getAuthUser() and Auth::login() from the updateLoginTime() method, which was re-authenticating the user with their full role permissions. It replaces this with a call to _refreshAuth(), which refreshes the authentication state while preserving the current (restricted) permission level associated with the API key, preventing privilege escalation.\u003c/p\u003e"
}
],
"value": "The fix removes the call to User::getAuthUser() and Auth::login() from the updateLoginTime() method, which was re-authenticating the user with their full role permissions. It replaces this with a call to _refreshAuth(), which refreshes the authentication state while preserving the current (restricted) permission level associated with the API key, preventing privilege escalation."
}
],
"title": "MISP Privilege Escalation: Read-Only API Key User Can Regain Full Role via updateLoginTime",
"x_gcve": [
{
"extensions": {
"bcp-05-x-01": {
"ai_annotations": [
{
"ai_level": "generated",
"description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
"gna_source": 1,
"models": [
{
"gna_source": 1,
"identifier": "qwen3.8:27b",
"name": "qwen3.8:27b",
"source": "ollama"
}
],
"review_status": "review",
"scope": "record",
"tags": [
"ai-computer-assisted:llm-generated",
"ai-computer-assisted:classification"
]
}
]
},
"bcp-05-x-02": {
"x_patch2vuln": {
"assumptions": [
"The tag_version_boundary indicates the fix commit is 141 commits before v2.5.47, implying v2.5.47 contains the fix; versions prior to v2.5.47 are assumed affected. No explicit version range is stated in the patch.",
"The _refreshAuth() method is assumed to preserve the current API-key-scoped permission level rather than re-fetching full role permissions; this is inferred from the commit subject and the removal of getAuthUser()/Auth::login().",
"The \u0027full role\u0027 referenced in the commit subject is assumed to be a higher-privilege role (e.g., admin or site admin) compared to the read-only API key scope.",
"CAPEC-114 is the closest available attack pattern; no CAPEC specifically models session privilege escalation via authentication refresh, so this mapping carries some uncertainty.",
"CVSS impact metrics (VC/VI/VA) assume the user\u0027s full role grants broad administrative access to the MISP instance; actual impact depends on the specific role assigned to the user."
],
"capecRationale": [
{
"capecId": "CAPEC-114",
"rationale": "The vulnerability arises because the updateLoginTime endpoint incorrectly reconfigures the user\u0027s privilege level during session refresh, upgrading from the API-key-scoped read-only permissions to the user\u0027s full role. The attacker exploits this misconfiguration by simply calling the endpoint. CAPEC-114 is the closest match as it covers scenarios where an application incorrectly assigns or maintains privilege levels. Uncertainty: no CAPEC specifically covers \u0027session privilege escalation via auth refresh,\u0027 so this is the best available approximation."
}
],
"commit": "fd27e592a617f21934f4a2c424e0578287affb36",
"confidence": "medium",
"credits": [
{
"lang": "en",
"type": "reporter",
"value": "Jeroen Pinoy"
},
{
"lang": "en",
"type": "remediation developer",
"value": "iglocska"
}
],
"cvssRationale": "AV:N - the endpoint is accessible over the network via HTTP POST. AC:L - the attack requires only a single POST request to a known endpoint; no race conditions or complex timing. AT:N - no manipulation of the target system is needed beyond the API call. PR:L - the attacker must possess a valid read-only API key (authenticated user). UI:N - no user interaction is required. VC:H - the escalated user can read all data in the MISP instance that their full role permits. VI:H - the escalated user can create, modify, and delete threat intelligence data. VA:H - the escalated user can delete data, modify configurations, or disrupt the MISP service. SC/SI/SA:N - the impact is confined to the MISP instance itself; no cascading impact on other systems is evidenced.",
"fixSummary": "The fix removes the call to User::getAuthUser() and Auth::login() from the updateLoginTime() method, which was re-authenticating the user with their full role permissions. It replaces this with a call to _refreshAuth(), which refreshes the authentication state while preserving the current (restricted) permission level associated with the API key, preventing privilege escalation.",
"generatedAt": "2026-09-21T12:37:21.609779Z",
"generator": "patch2vuln.py",
"model": "qwen3.8:27b",
"modelComparison": {
"rankings": [
{
"agreementScore": 9,
"assumptionCount": 5,
"confidence": "medium",
"model": "qwen3.8:27b",
"score": 5
}
],
"selectedModel": "qwen3.8:27b",
"selectionMethod": "deterministic-consensus-v1",
"selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
},
"patchSha256": "d4143841f5f0d52a14d8538760872a7d91f78ba110384eafa3eedcad8e6480d3",
"patchSummary": "In app/Controller/UsersController.php, the updateLoginTime() method is modified: the lines that fetched the user record via User::find(), called User::getAuthUser() to retrieve the full auth user, and called Auth::login() to re-establish the session are removed. These are replaced with a single call to $this-\u003e_refreshAuth(), which updates the authentication context without upgrading the permission scope. Net change: 1 insertion, 6 deletions.",
"patchTruncated": false,
"patches": [
{
"commit": "fd27e592a617f21934f4a2c424e0578287affb36",
"patchSha256": "d4143841f5f0d52a14d8538760872a7d91f78ba110384eafa3eedcad8e6480d3",
"source": "https://github.com/MISP/MISP/commit/fd27e592a.patch",
"sourceUrl": "https://github.com/MISP/MISP/commit/fd27e592a.patch",
"subject": "fix: [security] read only api keys can regain full role"
}
],
"source": "https://github.com/MISP/MISP/commit/fd27e592a.patch",
"subject": "fix: [security] read only api keys can regain full role",
"tagVersionBoundary": {
"commits_after_fix": 141,
"repository": "https://github.com/MISP/MISP",
"tag": "v2.5.47",
"version": "2.5.47",
"version_type": "semver"
},
"weaknessRationale": [
{
"cweId": "CWE-269",
"rationale": "The system fails to maintain the restricted privilege level (read-only) associated with the API key authentication. The updateLoginTime endpoint re-establishes the session with the user\u0027s full role permissions, effectively granting higher privileges than the authentication credential intended. This is a failure to properly manage and enforce the privilege boundary during session refresh."
}
]
}
}
},
"recordType": "advisory",
"vulnId": "GCVE-1-2026-20048"
}
]
}
},
"cveMetadata": {
"assignerOrgId": "00000000-0000-4000-9000-000000000000",
"cveId": "CVE-2026-94381",
"datePublished": "2026-09-21T12:42:44.209048Z",
"dateReserved": "2026-09-21T12:42:51.237Z",
"dateUpdated": "2026-09-21T12:42:51.306891Z",
"state": "PUBLISHED",
"vulnId": "GCVE-1-2026-20048"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.1"
}