The Agent Skills Directory

[SAFE]: The skill provides a robust framework for natural language orchestration with built-in safety constraints. It explicitly mandates human confirmation for any state-changing actions, such as modifying permissions or merging code, preventing autonomous or unauthorized changes to the environment.
[EXTERNAL_DOWNLOADS]: All external references are directed toward official GitHub documentation (docs.github.com) or the official GitHub blog. These are trusted sources and do not involve the execution of untrusted remote code.
[PROMPT_INJECTION]: While the skill ingests arbitrary user input to classify intent, it employs structured disambiguation logic and explicit confirmation steps, which mitigate the risk of indirect prompt injection influencing the resulting agent actions.

github-hub