codebase-advisor
Pass
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted content from external repositories, which presents a surface for indirect prompt injection attacks.
- Ingestion points: The agent reads the entire codebase, including source files, READMEs, ADRs, and configuration files, during the Recon and Audit phases as described in SKILL.md.
- Boundary markers: The skill contains explicit instructions in SKILL.md (Hard Rule 6) and references/closing-the-loop.md (Executor preamble) to treat repository content strictly as data and ignore any instructions found within, such as "ignore previous instructions."
- Capability inventory: The skill has access to tools for reading and searching files, creating GitHub issues (gh issue create), and spawning subagents for execution, as detailed in the allowed-tools and workflow sections of SKILL.md.
- Sanitization: There are mandatory rules in SKILL.md (Hard Rule 4) to never reproduce secret values and to report any detected injection attempts as security findings rather than following them.
Audit Metadata