grill-with-docs
Pass
Audited by Gen Agent Trust Hub on Jul 15, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses shell-level operations including
globandgrepto scan the codebase for existing features, route definitions, and shared types. These are standard operations for a development-focused tool. - [INDIRECT_PROMPT_INJECTION]: The skill represents a surface for indirect prompt injection because it ingests untrusted data from the local repository (markdown files, TypeScript source code) to inform its documentation updates.
- Ingestion points: Files located in
.agents/features/,packages/, anddocs/adr/are read via globbing and grep. - Boundary markers: The instructions do not define specific delimiters or security headers to differentiate between repository content and agent instructions.
- Capability inventory: The agent can read files and write to the local file system (updating
CONTEXT.md, creating ADRs). - Sanitization: No explicit sanitization or validation of the ingested file content is performed before it is used to generate or update documentation.
Audit Metadata