council
Pass
Audited by Gen Agent Trust Hub on Sep 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to gather external context such as codebase files, snippets, and metrics, creating a potential surface for indirect prompt injection if those files contain adversarial instructions.
- Ingestion points: Codebase files, snippets, issue text, and metrics gathered in Step 2.
- Boundary markers: The subagent prompt structure in Step 4 uses explicit 'Context:' headers to isolate external data from the primary instructions.
- Capability inventory: The skill can interact with tools like 'knowledge-ops', '/save-session', and update external issues on GitHub or Linear.
- Sanitization: The skill relies on structured prompting rather than explicit data sanitization or filtering logic.
Audit Metadata