ship-skill
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill performs several powerful shell operations including filesystem modification, Git commits/pushes to the user's repository, and global installation of software components using a dedicated CLI.
- [INDIRECT_PROMPT_INJECTION]: The skill acts as a pipeline for other skill files, creating an attack surface where malicious instructions in a processed skill could influence the agent's behavior during the 'Review' or 'Ship' phases.
- Ingestion points: The agent reads and processes markdown and configuration files from
~/dev/src/github.com/Zihad550/skills. - Boundary markers: No explicit XML delimiters or isolation instructions are provided when the agent reads the content of the skills being shipped.
- Capability inventory: The agent has access to
git,bash, and a global installation tool (skills add -g) that can modify system-wide agent configurations. - Sanitization: The skill includes a mandatory manual inspection step (Step 3) where the agent is explicitly instructed to scrub API keys, tokens, and other sensitive credentials from the files before staging them.
Audit Metadata