skills/kinantid/skills/refine/Gen Agent Trust Hub

refine

Pass

Audited by Gen Agent Trust Hub on Jul 19, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is designed to analyze and refine text based on communication principles. It operates entirely on local reference files and does not perform network operations or access sensitive system credentials.
  • [COMMAND_EXECUTION]: The skill package includes a local maintenance script scripts/check-principle-coverage.sh for developer validation of documentation completeness. The script uses standard system utilities and does not interact with external resources.
  • [PROMPT_INJECTION]: While the skill processes user-supplied text drafts, it has no capabilities (such as network or file writing) that could be exploited by an indirect prompt injection attack.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 19, 2026, 10:55 PM
Security Audit — agent-trust-hub — refine