The Agent Skills Directory

[SAFE]: The skill's behavior is consistent with its stated purpose of improving codebase architecture. It uses sub-agents for analysis and the gh tool for creating GitHub issues, which are standard operations for a developer-focused AI agent.
[SAFE]: The workflow includes a user checkpoint in Step 4 where the user selects or accepts a design before the GitHub issue is created in Step 5, ensuring human oversight before external write operations.
[SAFE]: No instances of prompt injection, data exfiltration, or obfuscation were found in the skill metadata or instructions. Remote code execution and persistence mechanisms are also absent.

improve-architecture