The Agent Skills Directory

[SAFE]: The skill implements a structured multi-step process for codebase analysis. It utilizes read-only search and diagnostics tools (ast-grep, LSP) to gather information from the local project environment.
[SAFE]: File writing functionality is restricted to the final step of the process and requires explicit user consent. The skill suggests a safe default directory (docs/ito-temp/) for saving generated documentation.
[SAFE]: No network-enabled tools or external data exfiltration patterns were detected. The skill operates entirely within the local codebase context.
[SAFE]: Prompt templates for sub-agents use clear markers to separate user input from system instructions, which helps mitigate basic direct prompt injection risks.

ito-explain