refactor-to-agenteq

Pass

Audited by Gen Agent Trust Hub on Sep 27, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPERSISTENCEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill uses npx to download and run the agenteq package and potentially hook managers like simple-git-hooks from the npm registry.
  • [REMOTE_CODE_EXECUTION]: Execution of third-party packages via npx (e.g., npx agenteq init, npx agenteq sync) involves running code from a remote registry.
  • [COMMAND_EXECUTION]: The skill executes several shell commands to manage the repository state, including:
  • git rm -r --cached to untrack files from the git index.
  • git status to verify repository state.
  • Commands to create or append to git hook files.
  • [PERSISTENCE]: The skill explicitly sets up automated execution via git hooks (post-checkout). It modifies or creates files in .husky/, lefthook.yml, package.json, or .git/hooks/post-checkout to ensure npx agenteq sync runs whenever a user switches branches.
  • [INDIRECT_PROMPT_INJECTION]: The skill has a large ingestion surface for untrusted data:
  • Ingestion points: It reads CLAUDE.md, .cursor/skills/, .mcp.json, and various other agent-specific configuration and guideline files.
  • Capability inventory: The skill has the ability to write to the file system, execute shell commands, and modify git configuration.
  • Sanitization: The skill does not explicitly describe sanitization for the content of the markdown files it refactors into the .ai/ directory.
  • Boundary markers: No explicit boundary markers are used when the agent reads or reconciles these files, though it does show diffs to the user in step 5.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 27, 2026, 10:07 AM