code-quality-trends
Pass
Audited by Gen Agent Trust Hub on Oct 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external sources which could contain malicious instructions designed to influence the agent's behavior during report generation.
- Ingestion points: The workflow ingests code search results (such as TODO, HACK, and FIXME strings) and GitHub Pull Request metadata, including titles, in steps 1, 3, and 4 of the SKILL.md file.
- Boundary markers: The skill instructions do not specify any delimiters or safety prompts to ensure the agent ignores potential instructions embedded within the ingested repository data.
- Capability inventory: The skill uses Sourcegraph search tools and the GitHub CLI (
gh) to read data and then generates a markdown report which includes a narrative summary written by the LLM. - Sanitization: There is no sanitization or filtering logic mentioned for the ingested PR titles or search results before they are processed by the LLM.
Audit Metadata