metaskill-authoring
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- SAFE (SAFE): The skill is a set of documentation and templates for authoring other skills. No security risks were identified.
- PROMPT_INJECTION (SAFE): No instructions attempt to override agent safety filters or system prompts.
- DATA_EXFILTRATION (SAFE): No file system access or external network calls are performed.
- REMOTE_CODE_EXECUTION (SAFE): No scripts or external dependencies are downloaded or executed.
- COMMAND_EXECUTION (SAFE): While the templates contain example commands (e.g.,
grep,npm test), these are illustrative placeholders within documentation and not executed by the skill itself.
Audit Metadata