setup-james-skills
Pass
Audited by Gen Agent Trust Hub on Nov 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes and modifies the root AGENTS.md file, which serves as a source of instructions for AI agents. This creates an attack surface where malicious instructions could be embedded in the data being reconciled.
- Ingestion points: The skill reads the repository's root AGENTS.md file.
- Boundary markers: The skill utilizes specific comment tags (james-skills:start and james-skills:end) to delimit its managed content block.
- Capability inventory: The skill performs file read and write operations on the AGENTS.md file.
- Sanitization: Instructions explicitly require preserving repository-specific instructions and unrelated content, and performing a diff to verify changes.
Audit Metadata