The Agent Skills Directory

[SAFE]: No malicious patterns or security risks were identified. The skill focuses on managing 'goal contracts' with comprehensive safety checks to prevent unauthorized execution of risky operations like production database migrations or authentication changes. The skill logic explicitly requires human confirmation for destructive, billing, or security-sensitive actions.
[COMMAND_EXECUTION]: The skill includes a test script scripts/test-lint-goal-prompt.py which uses subprocess.run to execute a local linter script (scripts/lint-goal-prompt.py) during its regression testing suite. This is a standard developer workflow for validating the quality of generated prompts and does not involve processing untrusted external data or arbitrary shell commands.

goal-gate