recipe-front-adjust
Pass
Audited by Gen Agent Trust Hub on Nov 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external sources and user-provided arguments.
- Ingestion points: External evidence fetched via the
external-resource-contextskill (Step 1) and the$ARGUMENTSparameter containing the user's adjustment request. - Boundary markers: Step 4 explicitly requires the agent to construct an adjustment context and wait for user confirmation (
[STOP]) before any file modifications occur. - Capability inventory: The skill possesses the capability to modify repository files (
Edit,Write,MultiEdit), invoke subagents (quality-fixer-frontend), and performgit commitoperations. - Sanitization: There is no automated sanitization or escaping of the input; the skill relies on the human-in-the-loop review of the context in Step 4 to identify and halt potentially malicious instructions embedded in the data.
Audit Metadata