AI Science Tracker

Taskflow Agent and CodeQL find code injection in PraisonAI GitHub Actions

PraisonAI claude-code-action workflow · AI involvement attributed to GitHub Security Lab Taskflow Agent and CodeQL

Finding and attribution

The reusable issue-triage action interpolated attacker-controlled issue bodies and titles into shell assignments in a workflow with access to an Anthropic API key, creating privileged code-execution and secret-exposure risk.

Product: PraisonAI claude-code-action workflow. AI system: GitHub Security Lab Taskflow Agent and CodeQL. Team / organization: GitHub Security Lab.

CVE identifiers: Not established in the cited evidence.

AI and human contribution (source account): GHSL explicitly credits both CodeQL and its Taskflow Agent with discovery; Peter Stöckli verified and reported the issue. The source does not apportion discovery between the two tools.

Human involvement is only described where the source account establishes it; unknown involvement does not mean none. Vendor confirmation of a flaw does not confirm every claim about AI contribution.

Discovery context: The public GHSL report describes the code-injection flaw and remediation; it does not claim exploitation in the wild.

Remediation status: fix documented. GHSL says the issue was fixed in a commit on 2025-12-16; the advisory does not identify a tagged release or establish whether downstream copies were updated.

A source-documented fix does not establish installed patches or current safety. Source checking does not independently reproduce the finding or establish observed exploitation.

Disclosed 16 Dec 2025. AI attribution dated 3 Apr 2026. Sources checked 14 Nov 2026; these are separate dates, not a live installation check.

Recorded CVSS: Not available · no assessment established here. Documented technical effects: code execution.

Severity, consequences and exploitation evidence

Case grouping uses the highest recorded CVSS score, with its version-specific severity band. It is descriptive, not a combined or editorial risk index. All recorded assessments remain below; scores from different versions and sources may disagree. CVSS specifications define version-specific scoring and severity bands.

No source-backed CVSS assessment is established in this catalog.

Case-level evidence: code execution.
Prerequisites / affected scope: Attacker-controlled workflow input can be injected into workflow code; source reports code injection..
Exploitation evidence: The cited source does not state whether this issue was exploited in the wild.. Impact source · checked 16 Nov 2026.

Documented fix: GHSL says the issue was fixed in a commit on 2025-12-16; the advisory does not identify a tagged release or establish whether downstream copies were updated.

Original evidence

Download this source record (JSON)

Open this case in the interactive tracker · Browse all static security cases