idea-refine
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill possesses a vulnerability surface for indirect prompt injection because it is designed to ingest and process data from the user's local codebase to provide context-aware ideation.
- Ingestion points: The
SKILL.mdfile instructs the agent to useGlob,Grep, andReadtools to scan the codebase for architecture, patterns, constraints, and precedents. - Boundary markers: No explicit delimiters, data-framing techniques, or instructions to disregard embedded commands are provided for the content read from the codebase.
- Capability inventory: The skill uses a local shell script (
scripts/idea-refine.sh) to create directories and has the ability to write markdown files todocs/ideas/. - Sanitization: There is no evidence of sanitization, filtering, or validation logic applied to the content ingested from the external files.
Audit Metadata