x9-skill-creator
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The validator utility (scripts/validate.py) and its test suite (scripts/test_validate.py) use subprocess.run to invoke system tools such as ruby, python3 -m py_compile, and bash -n. These operations are limited to performing static structural and syntax analysis on the files of the skill being audited.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted instructional content from external files. 1. Ingestion points: Reads SKILL.md and referenced files of target skills during Audit and Fix actions. 2. Boundary markers: The quality rubric (quality-rubric.md) defines explicit scope boundaries, authority limits, and preservation requirements. 3. Capability inventory: Limited to local file structural improvements and executing specific syntax checkers. 4. Sanitization: Relies on AI-driven auditing against the predefined rubric and structural validation checks for embedded secrets and junk files.
Audit Metadata