add-skill
Pass
Audited by Gen Agent Trust Hub on Jul 21, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to run local validation commands, such as
make checkorscripts/validate_skills.py, to ensure that newly created or modified skills meet repository standards. - [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it processes untrusted data from web searches and the local
.local/directory to inform the skill creation process. - Ingestion points: The agent reads the
.local/directory and performs web searches to gather facts and documentation. - Boundary markers: No specific delimiters are mandated for separating research data from instructions.
- Capability inventory: The skill can write files (SKILL.md, evals.json, README.md, CHANGELOG.md) and execute local shell commands (make check).
- Sanitization: The workflow relies on the agent's judgment to clean and verify research data before committing it to the repository.
Audit Metadata