The Agent Skills Directory

[SAFE]: The skill implements robust security guardrails for handling sensitive Git data. It explicitly forbids storing or printing credentials, private keys, and tokens, directing the agent to treat credential helpers as secret-handling configuration. Potential risks related to Indirect Prompt Injection and malicious execution are addressed through mandatory inspection steps for repository-controlled files like .gitattributes, .gitignore, and Git hooks. The instructions require that hooks and shell aliases be treated as policy-sensitive executable state, necessitating full disclosure of their executable and network effects before any modifications are made. No obfuscation, unauthorized network operations, or malicious patterns were detected.

configure-git