PYSEC-2026-1759
Vulnerability from pysec - Published: 2026-07-07 16:03 - Updated: 2026-07-07 17:25Affected Product: Parsl (Python Parallel Scripting Library)
Component: parsl.monitoring.visualization
Vulnerability Type: SQL Injection (CWE-89)
Severity: High (CVSS Rating Recommended: 7.5 - 8.6)
URL: https://github.com/Parsl/parsl/blob/master/parsl/monitoring/visualization/views.py
Summary
A SQL Injection vulnerability exists in the parsl-visualize component of the Parsl library. The application constructs SQL queries using unsafe string formatting (Python % operator) with user-supplied input (workflow_id) directly from URL routes. This allows an unauthenticated attacker with access to the visualization dashboard to inject arbitrary SQL commands, potentially leading to data exfiltration or denial of service against the monitoring database.
Root Cause Analysis
The vulnerability is located in parsl/monitoring/visualization/views.py. Multiple route handlers take the workflow_id parameter from the URL and inject it directly into a raw SQL query string without sanitization or parameterization.
Vulnerable Code Snippet 1
query = """SELECT task.task_id, task.task_func_name, task.task_depends, status.task_status_name
FROM task LEFT JOIN status
ON task.task_id = status.task_id
AND task.run_id = status.run_id
AND status.timestamp = (SELECT MAX(status.timestamp)
FROM status
WHERE status.task_id = task.task_id and status.run_id = task.run_id
)
WHERE task.run_id='%s'""" % (workflow_id)
Vulnerable Code Snippet 2
df_task_tries = pd.read_sql_query("""SELECT ...
WHERE task.task_id = try.task_id AND task.run_id='%s' and try.run_id='%s'"""
% (workflow_id, workflow_id), db.engine) # <--- Vulnerable
Impact
Data Exfiltration: An attacker can use UNION based injection to dump the entire contents of the monitoring database, including potentially sensitive environment variables, task parameters, or host information logged by the monitoring system.
Access Control Bypass: By injecting boolean logic (e.g., ' OR '1'='1), an attacker could bypass specific workflow filters to view data they are not authorized to see.
Denial of Service: Time-based attacks or resource-intensive queries (e.g., randomblob) could crash the visualization server or the database.
Proof of Concept (PoC)
Prerequisites:
Parsl installed with monitoring enabled (pip install 'parsl[monitoring,visualization]').
A running parsl-visualize server serving a database with at least one recorded workflow.
Reproduction Steps:
Identify a valid workflow ID (or use a known ID like default-run or a UUID).
Navigate to the dag_group_by_states endpoint using the following manipulated URLs to confirm SQL logic control (Boolean-based Blind SQLi).
Test 1: Boolean FALSE (Graph Disappears) Injecting a false condition (1=0) causes the query to return zero rows, resulting in an empty visualization.
http://<server>:8080/workflow/<VALID_ID>'%20AND%20'1'='0/dag_group_by_states
Test 2: Boolean TRUE (Graph Reappears) Injecting a true condition (1=1) restores the query logic, causing the graph to render correctly.
http://<server>:8080/workflow/<VALID_ID>'%20AND%20'1'='1/dag_group_by_states
| Name | purl | parsl | pkg:pypi/parsl |
|---|
{
"affected": [
{
"package": {
"ecosystem": "PyPI",
"name": "parsl",
"purl": "pkg:pypi/parsl"
},
"ranges": [
{
"events": [
{
"introduced": "0"
},
{
"fixed": "2026.01.05"
}
],
"type": "ECOSYSTEM"
}
],
"versions": [
"0.1.1",
"0.1.2",
"0.2.0",
"0.2.1",
"0.3.0",
"0.3.0a0",
"0.3.1",
"0.4.0",
"0.4.0a0",
"0.4.1",
"0.5.0",
"0.5.0a0",
"0.5.0a1",
"0.5.1",
"0.5.2",
"0.6.0",
"0.6.0a0",
"0.6.1",
"0.6.1a0",
"0.6.2a0",
"0.6.2a1",
"0.7.0",
"0.7.0a0",
"0.7.0a1",
"0.7.1",
"0.7.2",
"0.7.2a0",
"0.7.2a1",
"0.8.0",
"0.8.0a0",
"0.9.0",
"0.9.0a0",
"1.0.0",
"1.0.0a0",
"1.1.0",
"1.1.0a0",
"1.1.0a1",
"1.2.0",
"1.3.0.dev0",
"1.3.0.dev1",
"2022.10.10",
"2022.10.17",
"2022.10.24",
"2022.10.31",
"2022.10.4",
"2022.11.14",
"2022.11.21",
"2022.11.28",
"2022.11.7",
"2022.12.12",
"2022.12.19",
"2022.12.26",
"2022.12.5",
"2023.1.16",
"2023.1.2",
"2023.1.23",
"2023.1.30",
"2023.1.9",
"2023.10.16",
"2023.10.2",
"2023.10.23",
"2023.10.9",
"2023.11.13",
"2023.11.20",
"2023.11.27",
"2023.12.11",
"2023.12.18",
"2023.12.25",
"2023.12.4",
"2023.2.13",
"2023.2.20",
"2023.2.27",
"2023.2.6",
"2023.3.13",
"2023.3.20",
"2023.3.27",
"2023.3.6",
"2023.4.10",
"2023.4.17",
"2023.4.24",
"2023.4.3",
"2023.5.1",
"2023.5.15",
"2023.5.22",
"2023.5.29",
"2023.5.8",
"2023.6.12",
"2023.6.19",
"2023.6.26",
"2023.6.5",
"2023.7.10",
"2023.7.17",
"2023.7.24",
"2023.7.3",
"2023.7.31",
"2023.8.14",
"2023.8.21",
"2023.8.28",
"2023.8.7",
"2023.9.11",
"2023.9.18",
"2023.9.25",
"2023.9.4",
"2024.1.1",
"2024.1.15",
"2024.1.22",
"2024.1.29",
"2024.1.8",
"2024.10.14",
"2024.10.21",
"2024.10.28",
"2024.10.7",
"2024.11.11",
"2024.11.18",
"2024.11.25",
"2024.11.4",
"2024.12.16",
"2024.12.2",
"2024.12.23",
"2024.12.30",
"2024.12.9",
"2024.2.12",
"2024.2.19",
"2024.2.26",
"2024.2.5",
"2024.3.11",
"2024.3.18",
"2024.3.25",
"2024.3.4",
"2024.4.1",
"2024.4.15",
"2024.4.22",
"2024.4.29",
"2024.4.8",
"2024.5.13",
"2024.5.20",
"2024.5.27",
"2024.5.6",
"2024.6.10",
"2024.6.17",
"2024.6.24",
"2024.6.3",
"2024.7.1",
"2024.7.15",
"2024.7.22",
"2024.7.29",
"2024.7.8",
"2024.8.12",
"2024.8.19",
"2024.8.26",
"2024.8.5",
"2024.9.16",
"2024.9.2",
"2024.9.23",
"2024.9.30",
"2024.9.9",
"2025.1.13",
"2025.1.20",
"2025.1.27",
"2025.1.6",
"2025.10.13",
"2025.10.20",
"2025.10.27",
"2025.10.6",
"2025.11.10",
"2025.11.17",
"2025.11.24",
"2025.11.3",
"2025.12.1",
"2025.12.15",
"2025.12.22",
"2025.12.29",
"2025.12.8",
"2025.2.10",
"2025.2.17",
"2025.2.24",
"2025.2.3",
"2025.3.10",
"2025.3.17",
"2025.3.24",
"2025.3.3",
"2025.3.31",
"2025.4.14",
"2025.4.21",
"2025.4.28",
"2025.4.7",
"2025.5.12",
"2025.5.19",
"2025.5.26",
"2025.5.5",
"2025.6.16",
"2025.6.2",
"2025.6.23",
"2025.6.30",
"2025.6.9",
"2025.7.14",
"2025.7.21",
"2025.7.28",
"2025.7.7",
"2025.8.11",
"2025.8.18",
"2025.8.25",
"2025.8.4",
"2025.9.1",
"2025.9.15",
"2025.9.22",
"2025.9.29",
"2025.9.8"
]
}
],
"aliases": [
"CVE-2026-21892",
"GHSA-f2mf-q878-gh58"
],
"details": "**Affected Product:** Parsl (Python Parallel Scripting Library)\n\n**Component:** parsl.monitoring.visualization\n\n**Vulnerability Type:** SQL Injection (CWE-89)\n\n**Severity:** High (CVSS Rating Recommended: 7.5 - 8.6)\n\n**URL:** [https://github.com/Parsl/parsl/blob/master/parsl/monitoring/visualization/views.py]( https://github.com/Parsl/parsl/blob/master/parsl/monitoring/visualization/views.py)\n\n**Summary**\n\nA SQL Injection vulnerability exists in the parsl-visualize component of the Parsl library. The application constructs SQL queries using unsafe string formatting (Python % operator) with user-supplied input (workflow_id) directly from URL routes. This allows an unauthenticated attacker with access to the visualization dashboard to inject arbitrary SQL commands, potentially leading to data exfiltration or denial of service against the monitoring database.\n\n**Root Cause Analysis**\n\nThe vulnerability is located in parsl/monitoring/visualization/views.py. Multiple route handlers take the workflow_id parameter from the URL and inject it directly into a raw SQL query string without sanitization or parameterization.\n\nVulnerable Code Snippet 1 \n\n```\n query = \"\"\"SELECT task.task_id, task.task_func_name, task.task_depends, status.task_status_name\n FROM task LEFT JOIN status\n ON task.task_id = status.task_id\n AND task.run_id = status.run_id\n AND status.timestamp = (SELECT MAX(status.timestamp)\n FROM status\n WHERE status.task_id = task.task_id and status.run_id = task.run_id\n )\n WHERE task.run_id=\u0027%s\u0027\"\"\" % (workflow_id)\n```\n\nVulnerable Code Snippet 2\n```\ndf_task_tries = pd.read_sql_query(\"\"\"SELECT ...\n WHERE task.task_id = try.task_id AND task.run_id=\u0027%s\u0027 and try.run_id=\u0027%s\u0027\"\"\"\n % (workflow_id, workflow_id), db.engine) # \u003c--- Vulnerable \n```\n\n**Impact**\n\n**Data Exfiltration:** An attacker can use UNION based injection to dump the entire contents of the monitoring database, including potentially sensitive environment variables, task parameters, or host information logged by the monitoring system.\n\n**Access Control Bypass:** By injecting boolean logic (e.g., \u0027 OR \u00271\u0027=\u00271), an attacker could bypass specific workflow filters to view data they are not authorized to see.\n\n**Denial of Service:** Time-based attacks or resource-intensive queries (e.g., randomblob) could crash the visualization server or the database.\n\n**Proof of Concept (PoC)**\n\n**Prerequisites:**\n\nParsl installed with monitoring enabled (pip install \u0027parsl[monitoring,visualization]\u0027).\n\nA running parsl-visualize server serving a database with at least one recorded workflow.\n\n**Reproduction Steps:**\n\nIdentify a valid workflow ID (or use a known ID like default-run or a UUID).\n\nNavigate to the dag_group_by_states endpoint using the following manipulated URLs to confirm SQL logic control (Boolean-based Blind SQLi).\n\nTest 1: Boolean FALSE (Graph Disappears)\nInjecting a false condition (1=0) causes the query to return zero rows, resulting in an empty visualization.\n\n`http://\u003cserver\u003e:8080/workflow/\u003cVALID_ID\u003e\u0027%20AND%20\u00271\u0027=\u00270/dag_group_by_states\n`\n\nTest 2: Boolean TRUE (Graph Reappears)\nInjecting a true condition (1=1) restores the query logic, causing the graph to render correctly.\n\n`http://\u003cserver\u003e:8080/workflow/\u003cVALID_ID\u003e\u0027%20AND%20\u00271\u0027=\u00271/dag_group_by_states\n`",
"id": "PYSEC-2026-1759",
"modified": "2026-07-07T17:25:00.056998Z",
"published": "2026-07-07T16:03:16.784288Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/Parsl/parsl/security/advisories/GHSA-f2mf-q878-gh58"
},
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-21892"
},
{
"type": "WEB",
"url": "https://github.com/Parsl/parsl/commit/013a928461e70f38a33258bd525a351ed828e974"
},
{
"type": "PACKAGE",
"url": "https://github.com/Parsl/parsl"
},
{
"type": "PACKAGE",
"url": "https://pypi.org/project/parsl"
},
{
"type": "ADVISORY",
"url": "https://github.com/advisories/GHSA-f2mf-q878-gh58"
}
],
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:L",
"type": "CVSS_V3"
}
],
"summary": "Parsl Monitoring Visualization Vulnerable to SQL Injection"
}
Drag nodes to arrange them. Double-click a related vulnerability to open it, or a source to visit its link. The Physics button on the left tunes the layout and the simulation; the legend at the bottom left hides or shows nodes by kind and by sighting type.
Sightings
| Author | Source | Type | Date | Other |
|---|
Nomenclature
- Seen: The vulnerability was mentioned, discussed, or observed by the user.
- Confirmed: The vulnerability has been validated from an analyst's perspective.
- Published Proof of Concept: A public proof of concept is available for this vulnerability.
- Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
- Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
- Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
- Not confirmed: The user expressed doubt about the validity of the vulnerability.
- Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.
Vulnerability timeline
Sightings and recorded dates, grouped by day · UTC
How durations are calculated
Following the VTEM framework (Black Hat USA 2015), time to patch starts when the vendor or a coordinator was informed, total time to patch at discovery, and disclosure to patch at public disclosure; vendor response time runs from the report to its acknowledgement. Time to exploit and time unpatched start at the earliest discovery or report, falling back to public disclosure. Unpatched time ends at the first recorded patch, or the calculation time when no patch is recorded. Exploitation to patch shows the interval between those two events; a published proof of concept is a public exploit, not exploitation.
Each result shows its source dates. A negative interval means the end event came first; when date-only inputs leave the order open, the result says so. Date-only inputs are approximate. The earliest evidence wins: KEV report dates and sightings are labelled as such. Records describing another vulnerability are listed but never calculated from, and a PoC alone does not establish exploitation. Missing patch evidence does not prove that no patch exists. Filters below change the event list only.
Loading timeline…
The approach is described in our paper Mapping CVEs to MITRE ATT&CK Techniques: A Curated Gold-Set Classifier and the Limits of LLM-Assisted Label Expansion.
Browse all ATT&CK techniques and the vulnerabilities related to each.
Related by attack behaviour
Vulnerabilities whose description is nearest to this one in the vector space of the CIRCL/vulnerability-attack-technique-biencoder model. This is a similarity search over the bi-encoder space (plain cosine), not a classification, and it has no measured accuracy.