command-skill-creator
Pass
Audited by Gen Agent Trust Hub on May 5, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements a security-first design philosophy by mandating an 'Audit Checklist' for all generated commands. This includes a requirement to set
disable-model-invocation: truefor any command with side effects, ensuring they only run upon explicit user request. - [SAFE]: It promotes the use of 'Approval Gates' (e.g., 'STOP and wait for user approval') before irreversible or destructive actions like commits, deploys, or file deletions.
- [COMMAND_EXECUTION]: The skill provides templates and patterns for executing common developer tasks using shell commands (e.g., git, pnpm, docker, kubectl). These are intended for explicit automation and are restricted by the platform's tool-calling policies.
- [SAFE]: The instructions explicitly forbid hardcoded absolute paths, instead requiring the use of relative paths or environment variables like
${CLAUDE_PROJECT_DIR}, which prevents accidental exposure of host-specific file structures.
Audit Metadata