meta-skill-creator
Pass
Audited by Gen Agent Trust Hub on Jul 16, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes the Bash tool to execute internal Python validation scripts (e.g., scripts/check_meta_skill_package.py). These scripts perform deterministic checks on file existence, content phrases, and cryptographic hashes to verify the consistency of the skill package.
- [SAFE]: The skill's core contract and reference documentation explicitly forbid the inclusion of sensitive data, credentials, or private machine paths in distributed skill packages, demonstrating an awareness for data protection in generated content.
- [DATA_EXFILTRATION]: While the skill mentions the possibility of online evidence fetching for research, the provided code artifacts contain no networking logic or hidden exfiltration channels. All script operations are strictly localized to the project and 'acceptance run' directories.
- [PROMPT_INJECTION]: Instructions within the skill define how the agent should prioritize this framework for skill-related tasks. These are benign operational directives and do not contain patterns designed to override safety guardrails or bypass system-level instructions.
Audit Metadata