triage-findings
Pass
Audited by Gen Agent Trust Hub on Sep 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted data from local files and remote GitHub PRs, which serves as an entry point for instructions that could influence the agent's triage decisions.
- Ingestion points: Instructions in
SKILL.mddirect the agent to load the active findings record frompath/findings.mdand context from an optional GitHub PRurlprovided in the prompt. - Boundary markers: The instructions do not define specific delimiters or "ignore embedded instructions" warnings for the content read from findings files or PR URLs.
- Capability inventory: The results of the triage phase are used to decide the execution of subsequent tools, including
scan-codebase,test-codebase,sync-findings, andresolve-findings. - Sanitization: There is no explicit logic for validating or escaping the content of the findings records before the agent interpolates the data into its decision-making context.
Audit Metadata