ci-failure-triage
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's primary function involves analyzing external, untrusted data including runner logs, stack traces, and environment metadata. An adversary could attempt to manipulate the triage verdict by embedding deceptive signals (e.g., specific crash signatures) within the test output. \n
- Ingestion points: Test logs, stack traces, run history, and environment metadata extracted in Step 1 (SKILL.md). \n
- Boundary markers: The skill does not define specific delimiters or instructions for the agent to distinguish log content from its own decision logic. \n
- Capability inventory: The skill is restricted to logical classification and text emission; it has no capabilities for tool invocation, network requests, or persistent file system modifications. \n
- Sanitization: The instructions do not require validation or sanitization of the input log data. \n- [SAFE]: All external links and documentation references target trusted organizations (Google, Microsoft/Playwright) or well-known technical standards (W3C, ISTQB). The logic is transparent and does not involve hidden command execution or data exfiltration.
Audit Metadata