create-skill
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to use standard Unix utilities (mkdir, ls, grep, wc, find, ln) for project organization and verification. These actions are strictly limited to the local file system and repository structure.
- [DATA_EXPOSURE]: The skill involves creating symlinks within the user's home directory (~/.claude/skills/) for agent platform integration. This is a documented and standard practice for the target environment and does not involve unauthorized data access.
- [REMOTE_CODE_EXECUTION]: The skill references local scripts (sync-discovery-symlinks.sh) and npm-based tasks (translate:scaffold, coverage:evals) for repository maintenance. No patterns of downloading and executing remote code from untrusted external sources were detected.
Audit Metadata