skill-harmonization
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill requires and executes local CLI tools to perform its tasks. Evidence includes calls to git rev-parse --show-toplevel, ai-skillet map, and ai-skillet doctor.
- [INDIRECT_PROMPT_INJECTION]: The skill inventories and analyzes external files, creating a surface where instructions embedded in those files could influence agent behavior. 1. Ingestion points: The agent reads SKILL.md files, AGENTS.md, CLAUDE.md, and referenced automation scripts from the repository. 2. Boundary markers: The instructions command the agent to separate observed facts from inferred design intent and restrict evidence gathering to specific skill-related files. 3. Capability inventory: The agent has the authority to execute CLI tools and perform file writes during implementation tasks. 4. Sanitization: No explicit programmatic sanitization is defined; the skill relies on the agent's adherence to instructions to distinguish data from instructions.
Audit Metadata