The Agent Skills Directory

[SAFE]: The skill operates as a local diagnostic utility. It reads project files and error logs to match them against a hardcoded catalogue of failures. It does not perform network exfiltration, execute remote code, or attempt to bypass security guardrails.
[SAFE]: The skill provides explicit 'Stop and ask a human' rules for sensitive operations, including handling passwords, sudo commands, signing keys, and license agreements, ensuring the agent does not perform high-risk actions autonomously.
[SAFE]: The scripts are built using standard Node.js built-in modules and follow a zero-dependency approach, which minimizes the risk of supply chain attacks.

troubleshooting-playbook