init-jean-json

Pass

Audited by Gen Agent Trust Hub on Jul 27, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses git worktree and zsh to execute dynamically generated setup scripts in a test environment to verify the onboarding manifest. This involves shell commands for exploration, installation, and environment validation.
  • [EXTERNAL_DOWNLOADS]: The skill instructs the agent to download project dependencies and Docker images from well-known services and registries (NPM, PyPI, Docker Hub, etc.) during the setup phase.
  • [PROMPT_INJECTION]: Indirect injection surface identified as the skill reads and acts upon content from repository files like package.json, jean.json, and AGENTS.md. A malicious repository could attempt to influence the generated setup scripts.
  • Ingestion points: package.json, .gitignore, jean.json, .worktreeinclude, AGENTS.md, CLAUDE.md.
  • Boundary markers: None present in the instructions.
  • Capability inventory: git worktree management, shell execution (zsh -l -c), file operations (cp, rm), and package manager execution.
  • Sanitization: No explicit sanitization of repository content is required before use in generated scripts.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 27, 2026, 12:27 PM
Security Audit — agent-trust-hub — init-jean-json