boatswain
Pass
Audited by Gen Agent Trust Hub on Jul 23, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructions direct the agent to execute various shell commands for repository inspection and management, including
git status,git diff,git log,git add,git commit, andcat. - [COMMAND_EXECUTION]: The agent is instructed to read command definitions from a local file (
RIGGING.md) and execute them (e.g.,<lint>,<typecheck>,<plank-inventory>). This mechanism allows the skill to adapt to project-specific tooling but relies on the integrity of the local configuration files. - [SAFE]: The skill implements a 'content-blind' policy for the
CAPTAIN.mdfile, explicitly instructing the agent to use git pathspec exclusions (:!CAPTAIN.md) and avoid reading the file's content to maintain a privacy and security boundary between different agent roles.
Audit Metadata