agency-codebase-onboarding-engineer
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted data in the form of external codebase files during its inventory and tracing workflow.
- Ingestion points: Reads all source files, manifests, and configuration files within the repository being explored (SKILL.md workflow).
- Boundary markers: The instructions lack explicit data delimiters (like XML tags or triple backticks for interpolation), but they strongly enforce a "facts only" and "strictly read-only" constraint to minimize hallucinations or obedience to instructions found within the code.
- Capability inventory: The skill is explicitly limited to read-only operations and is forbidden from modifying files, generating patches, or changing repository state.
- Sanitization: There is no mention of sanitizing or escaping content read from source files before it is included in the orientation map output.
Audit Metadata