grill-with-docs

Pass

Audited by Gen Agent Trust Hub on May 17, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill is designed for documentation management and terminology refinement. It interacts with the local filesystem to read code and write documentation updates, which is standard for developer-centric AI skills. No malicious activity, such as data exfiltration or unauthorized command execution, was identified.\n- [PROMPT_INJECTION]: The skill incorporates a surface for indirect prompt injection as it reads and processes external data from the project's source code and documentation to inform its updates. While this is a common feature for code-analysis tools, it lacks specific sanitization or boundary measures to isolate potentially untrusted content from the agent's instructions.\n
  • Ingestion points: Reads project files including CONTEXT.md, CONTEXT-MAP.md, ADR files in docs/adr/, and the src/ directory (SKILL.md).\n
  • Boundary markers: The instructions do not specify the use of clear delimiters or instructions to ignore embedded commands when processing codebase content.\n
  • Capability inventory: The skill uses tools to explore the codebase (file search/read) and tools to write or update documentation files (SKILL.md).\n
  • Sanitization: There are no explicit steps defined to sanitize, escape, or validate the content retrieved from the file system before it is processed or written back to the documentation.
Audit Metadata
Risk Level
SAFE
Analyzed
May 17, 2026, 01:46 PM
Security Audit — agent-trust-hub — grill-with-docs