The Agent Skills Directory

[SAFE]: The skill exhibits no malicious behavior and follows industry best practices for autonomous agent orchestration. It utilizes structured audit trails and persistent state management to ensure transparency and recoverability across sessions.
[COMMAND_EXECUTION]: The orchestrator performs legitimate development-related commands such as git (for worktree management and branch operations), cargo-mutants (for quality gating), and execution of local test suites. These actions are within the expected scope of a build orchestrator and are limited to the project directory.
[EXTERNAL_DOWNLOADS]: The skill's documentation provides clear instructions for the user to manually install dependencies from the author's own repository (jwilger/agent-skills) using the npx skills add utility. This is a transparent and standard installation pattern.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input from event models and CI outputs. Potential risks are effectively mitigated by a mandatory human approval gate for the 'Agent Delivery Contract' and a multi-agent ensemble review process that validates all specifications before implementation begins.

pipeline