competition-submission-checker

Pass

Audited by Gen Agent Trust Hub on Jul 15, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill's primary function is to provide a validation checklist for CSV submissions and notebook outputs. It does not contain instructions for performing high-risk operations or accessing sensitive user data outside of the competition directory.
  • [DATA_EXFILTRATION]: The skill includes explicit defensive instructions under the 'Safety' section, forbidding the agent from using git, GitHub, or Google services, and strictly prohibiting external file submission without explicit user consent.
  • [COMMAND_EXECUTION]: References a local maintenance script scripts/export-gemini-skill.py for skill portability. This is documented as a manual rebuild step for the user and is not invoked by the agent during the automated validation workflow.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 15, 2026, 06:06 PM
Security Audit — agent-trust-hub — competition-submission-checker