The Agent Skills Directory

[SAFE]: The skill instructions provide a structured approach for an AI agent to act as a coordinator. It includes specific safety guidelines, such as 'Never merge, close PRs, comment, label, deploy, force-push, or touch production/secrets unless the start contract grants that exact action.'
[COMMAND_EXECUTION]: The skill suggests using npx skills for discovery. npx is a standard Node.js utility used here for exploring available capabilities within the environment, which is expected behavior for an orchestration tool.
[SAFE]: While the skill interacts with external data (PR metadata, bot reviews, and user messages), it acts as a coordinator with a 'human-in-the-loop' requirement for approval gates and final decisions, mitigating risks associated with indirect prompt injection.

workflow