skills/dmytri/shipshape/bosun/Gen Agent Trust Hub

bosun

Pass

Audited by Gen Agent Trust Hub on Jun 29, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill uses npx to execute tools such as gplint and cucumber-js. These downloads originate from the npm registry, which is an established and well-known service for Node.js development.
  • [COMMAND_EXECUTION]: Performs automated repository management using git (status, diff, log, commit) and grep for pattern matching. It also executes a 'discover' command defined in the project's RIGGING.md file.
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it ingests and executes commands specified in external project files like RIGGING.md and processes data from CAPTAIN.md and watchbill.json. 1. Ingestion points: RIGGING.md (tooling values), CAPTAIN.md (spec quality), watchbill.json (work selection). 2. Boundary markers: Absent. 3. Capability inventory: File system modification (annotations, deletions), local git commits, and shell command execution via npx and git. 4. Sanitization: Absent.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 29, 2026, 02:13 PM
Security Audit — agent-trust-hub — bosun