Architect Engineer — World-Class AI Build System

This skill is internally coherent as a comprehensive engineering playbook, but it is deliberately aggressive: it tells the agent to act as a production-ready architect and to 'trigger aggressively' for any technical task. If you install or enable it, consider these precautions: - Do not provide production credentials or broad-scoped secrets to the agent. If testing, use sandboxed or low-permission API keys and throwaway accounts. - Expect the skill to ask for configuration (.env files, API keys, repo access). Only grant the minimum scope necessary (principle of least privilege) and prefer service accounts with restricted permissions. - The bundle contains company-specific guidance (emails, Replit workflows, and commit policies). Make sure any automatic commit/push steps are reviewed and that you understand where code would be stored or pushed. - If you need to limit autonomous actions, consider disabling autonomous invocation for this skill or require explicit user consent before it performs external operations (git pushes, deployment, API calls). - Review the reference files that the skill will use (they're bundled) so you understand what patterns and outputs it will produce — especially anything that constructs or stores tokens, writes .env files, or contains instructions to ping external healthcheck endpoints. If you want to proceed: test it in an isolated environment first (local repo, sandbox API keys, ephemeral resources), and audit any outputs and requests it generates before giving access to production systems.