GHSA-228V-WC5R-J8M7
Vulnerability from github – Published: 2026-03-12 14:20 – Updated: 2026-03-12 14:21Summary
OliveTin’s live EventStream broadcasts execution events and action output to authenticated dashboard subscribers without enforcing per-action authorization. A low-privileged authenticated user can receive output from actions they are not allowed to view, resulting in broken access control and sensitive information disclosure. I validated this on OliveTin 3000.10.2.
Details
The issue is in the live event streaming path.
EventStream() only checks whether the caller may access the dashboard, then registers the user as a stream subscriber:
- service/internal/api/api.go:776
After subscription, execution events are broadcast to all connected clients without checking whether each recipient is authorized to view logs for the action:
- service/internal/api/api.go:846 OnExecutionStarted
- service/internal/api/api.go:869 OnExecutionFinished
- service/internal/api/api.go:1047 OnOutputChunk
The event payload includes action output through:
- service/internal/api/api.go:295 internalLogEntryToPb
- service/internal/api/api.go:302 Output
By contrast, the normal log APIs do apply per-action authorization checks:
- service/internal/api/api.go:518 GetLogs
- service/internal/api/api.go:585 GetActionLogs
- service/internal/api/api.go:544 isLogEntryAllowed
Root cause:
- the subscription path enforces only coarse dashboard access
- execution callbacks broadcast to every connected client
- no per-recipient ACL check is applied before sending action metadata or output
I validated the issue using:
- an admin user with full ACLs
- an alice user with no ACLs
- a protected action that outputs TOPSECRET=alpha-bravo-charlie
Despite having no relevant ACLs, alice still receives the ExecutionFinished event for the privileged action, including the protected output.
PoC
Tested version:
- 3000.10.2
- Fetch and check out 3000.10.2 in a clean worktree:
git -C OliveTin fetch origin tag 3000.10.2
git -C OliveTin worktree add /home/kali/CVE/OliveTin-3000.10.2 3000.10.2
- Copy the PoC test into the clean tree:
cp OliveTin/service/internal/api/event_stream_leak_test.go \
OliveTin-3000.10.2/service/internal/api/
- Run the targeted PoC test:
cd OliveTin-3000.10.2/service
go test ./internal/api -run TestEventStreamLeaksUnauthorizedExecutionOutput -count=1 -timeout 30s -v
- Optional: save validation output:
go test ./internal/api -run TestEventStreamLeaksUnauthorizedExecutionOutput -count=1 -timeout 30s -v \
2>&1 | tee /tmp/olivetin_eventstream_3000.10.2.log
Observed validation output:
=== RUN TestEventStreamLeaksUnauthorizedExecutionOutput
time="2026-03-01T04:44:59-05:00" level=info msg="Action requested" actionTitle=secret-action tags="[]"
time="2026-03-01T04:44:59-05:00" level=info msg="Action parse args - Before" actionTitle=secret-action cmd="echo 'TOPSECRET=alpha-bravo-charlie'"
time="2026-03-01T04:44:59-05:00" level=info msg="Action parse args - After" actionTitle=secret-action cmd="echo 'TOPSECRET=alpha-bravo-charlie'"
time="2026-03-01T04:44:59-05:00" level=info msg="Action started" actionTitle=secret-action timeout=1
time="2026-03-01T04:44:59-05:00" level=info msg="Action finished" actionTitle=secret-action exit=0 outputLength=30 timedOut=false
--- PASS: TestEventStreamLeaksUnauthorizedExecutionOutput (0.00s)
PASS
ok github.com/OliveTin/OliveTin/internal/api 0.025s
What this proves:
- admin can execute the protected action
- alice has no ACLs
- alice still receives the streamed completion event for the protected action
- protected action output is exposed through the event stream
Impact
This is an authenticated broken access control / information disclosure vulnerability.
A low-privileged authenticated user can subscribe to EventStream and receive:
- action execution metadata
- execution tracking IDs
- initiating username
- live output chunks
- final command output
Who is impacted:
- multi-user OliveTin deployments
- environments where privileged actions produce secrets, tokens, internal system details, or other sensitive operational output
- deployments where lower-privileged authenticated users can access the dashboard and subscribe to live events
This bypasses intended per-action log/view restrictions for protected actions.
{
"affected": [
{
"package": {
"ecosystem": "Go",
"name": "github.com/OliveTin/OliveTin"
},
"ranges": [
{
"events": [
{
"introduced": "0"
},
{
"fixed": "3000.10.2"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [
"CVE-2026-32102"
],
"database_specific": {
"cwe_ids": [
"CWE-284",
"CWE-863"
],
"github_reviewed": true,
"github_reviewed_at": "2026-03-12T14:20:16Z",
"nvd_published_at": "2026-03-11T21:16:16Z",
"severity": "HIGH"
},
"details": "### Summary\n\n OliveTin\u2019s live EventStream broadcasts execution events and action output to authenticated dashboard subscribers without enforcing per-action authorization. A low-privileged authenticated user can receive output from actions they are\n not allowed to view, resulting in broken access control and sensitive information disclosure. I validated this on OliveTin 3000.10.2.\n\n\n\n\n### Details\nThe issue is in the live event streaming path.\n\n EventStream() only checks whether the caller may access the dashboard, then registers the user as a stream subscriber:\n\n - service/internal/api/api.go:776\n\n After subscription, execution events are broadcast to all connected clients without checking whether each recipient is authorized to view logs for the action:\n\n - service/internal/api/api.go:846 OnExecutionStarted\n - service/internal/api/api.go:869 OnExecutionFinished\n - service/internal/api/api.go:1047 OnOutputChunk\n\n The event payload includes action output through:\n\n - service/internal/api/api.go:295 internalLogEntryToPb\n - service/internal/api/api.go:302 Output\n\n By contrast, the normal log APIs do apply per-action authorization checks:\n\n - service/internal/api/api.go:518 GetLogs\n - service/internal/api/api.go:585 GetActionLogs\n - service/internal/api/api.go:544 isLogEntryAllowed\n\n Root cause:\n\n - the subscription path enforces only coarse dashboard access\n - execution callbacks broadcast to every connected client\n - no per-recipient ACL check is applied before sending action metadata or output\n\n I validated the issue using:\n\n - an admin user with full ACLs\n - an alice user with no ACLs\n - a protected action that outputs TOPSECRET=alpha-bravo-charlie\n\n Despite having no relevant ACLs, alice still receives the ExecutionFinished event for the privileged action, including the protected output.\n\n\n\n### PoC\nTested version:\n```\n - 3000.10.2\n```\n 1. Fetch and check out 3000.10.2 in a clean worktree:\n```bash\n git -C OliveTin fetch origin tag 3000.10.2\n git -C OliveTin worktree add /home/kali/CVE/OliveTin-3000.10.2 3000.10.2\n```\n 2. Copy the PoC test into the clean tree:\n```bash\n cp OliveTin/service/internal/api/event_stream_leak_test.go \\\n OliveTin-3000.10.2/service/internal/api/\n```\n 3. Run the targeted PoC test:\n```bash\n cd OliveTin-3000.10.2/service\n go test ./internal/api -run TestEventStreamLeaksUnauthorizedExecutionOutput -count=1 -timeout 30s -v\n```\n 4. Optional: save validation output:\n```bash\n go test ./internal/api -run TestEventStreamLeaksUnauthorizedExecutionOutput -count=1 -timeout 30s -v \\\n 2\u003e\u00261 | tee /tmp/olivetin_eventstream_3000.10.2.log\n```\n Observed validation output:\n```bash\n === RUN TestEventStreamLeaksUnauthorizedExecutionOutput\n time=\"2026-03-01T04:44:59-05:00\" level=info msg=\"Action requested\" actionTitle=secret-action tags=\"[]\"\n time=\"2026-03-01T04:44:59-05:00\" level=info msg=\"Action parse args - Before\" actionTitle=secret-action cmd=\"echo \u0027TOPSECRET=alpha-bravo-charlie\u0027\"\n time=\"2026-03-01T04:44:59-05:00\" level=info msg=\"Action parse args - After\" actionTitle=secret-action cmd=\"echo \u0027TOPSECRET=alpha-bravo-charlie\u0027\"\n time=\"2026-03-01T04:44:59-05:00\" level=info msg=\"Action started\" actionTitle=secret-action timeout=1\n time=\"2026-03-01T04:44:59-05:00\" level=info msg=\"Action finished\" actionTitle=secret-action exit=0 outputLength=30 timedOut=false\n --- PASS: TestEventStreamLeaksUnauthorizedExecutionOutput (0.00s)\n PASS\n ok github.com/OliveTin/OliveTin/internal/api 0.025s\n```\n What this proves:\n\n - admin can execute the protected action\n - alice has no ACLs\n - alice still receives the streamed completion event for the protected action\n - protected action output is exposed through the event stream\n\n\n### Impact\n This is an authenticated broken access control / information disclosure vulnerability.\n\n A low-privileged authenticated user can subscribe to EventStream and receive:\n\n - action execution metadata\n - execution tracking IDs\n - initiating username\n - live output chunks\n - final command output\n\n Who is impacted:\n\n - multi-user OliveTin deployments\n - environments where privileged actions produce secrets, tokens, internal system details, or other sensitive operational output\n - deployments where lower-privileged authenticated users can access the dashboard and subscribe to live events\n\n This bypasses intended per-action log/view restrictions for protected actions.",
"id": "GHSA-228v-wc5r-j8m7",
"modified": "2026-03-12T14:21:00Z",
"published": "2026-03-12T14:20:16Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/OliveTin/OliveTin/security/advisories/GHSA-228v-wc5r-j8m7"
},
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-32102"
},
{
"type": "PACKAGE",
"url": "https://github.com/OliveTin/OliveTin"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N",
"type": "CVSS_V4"
}
],
"summary": "OliveTin Vulnerable to Unauthorized Action Output Disclosure via EventStream"
}
Sightings
| Author | Source | Type | Date |
|---|
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.