grill-with-docs

Pass

Audited by Gen Agent Trust Hub on Aug 28, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill does not contain any detected malicious patterns. It focuses on project planning and documentation maintenance by exploring existing files and interacting with the user.\n- [INDIRECT_PROMPT_INJECTION]: While the skill ingests codebase content, which is a surface for indirect prompt injection, this behavior is central to its primary purpose of documentation synchronization and plan verification.\n
  • Ingestion points: The agent is instructed to explore the codebase, including src/, CONTEXT.md, and ADR files (SKILL.md line 18, 26-44).\n
  • Boundary markers: The instructions do not define delimiters or safety markers for separating codebase content from agent instructions.\n
  • Capability inventory: The skill is authorized to update CONTEXT.md and create/edit ADR files in docs/adr/ (SKILL.md line 66, 80).\n
  • Sanitization: There is no mention of sanitizing content extracted from the codebase before it is used to update project documentation.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 28, 2026, 12:25 AM
Security Audit — agent-trust-hub — grill-with-docs