skills/byronfinn/dev-skills/grill/Gen Agent Trust Hub

grill

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill facilitates a structured workflow for technical planning and documentation updates. All identified file operations, including reading PRD files, updating glossaries in CONTEXT.md, and generating ADRs, are confined to the repository's documentation directories and represent standard development practices.
  • [INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it is instructed to process untrusted data from the codebase and project requirement documents. However, this is inherent to its primary function and lacks high-risk capabilities.
  • Ingestion points: The skill reads docs/prd/PRD-NNNN-<title>.md, CONTEXT.md, and the project's source code files.
  • Boundary markers: The instructions define a strict one-question-per-message interview loop, though it does not explicitly use sanitization delimiters for the content it reads.
  • Capability inventory: The skill can read and update repository files and interact with the project's issue tracker to sync parent issues.
  • Sanitization: No explicit sanitization or filtering of the content read from external files is described in the logic.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 06:22 AM
Security Audit — agent-trust-hub — grill