The Agent Skills Directory

[SAFE]: The skill utilizes standard command-line tools like git and grep to perform local context analysis, identifying symbol usage and dependency relationships without external network access.
[SAFE]: The workflow incorporates a mandatory security self-check phase (Step 4) that specifically instructs the agent to verify that no API keys, tokens, or PII are inadvertently included in the generated documentation.
[SAFE]: Instruction design focuses exclusively on documentation and metadata, explicitly forbidding changes to program logic or the modification of executable code segments.

comment-code