The Agent Skills Directory

[COMMAND_EXECUTION]: The skill implements subprocess execution patterns using the Python subprocess module to interact with the GitHub CLI (gh). It enforces security best practices by requiring argument arrays and setting shell=False, effectively preventing shell-based command injection (CWE-78).
[CREDENTIALS_UNSAFE]: Authentication patterns prioritize security by fetching sensitive tokens from environment variables rather than hardcoding them. The provided templates include logic to validate token formats before use, reducing the risk of accidental exposure or misuse.
[SAFE]: The skill serves as a security-focused knowledge base, incorporating defensive coding techniques like retry logic with exponential backoff, rate limiting, and whitelisting of allowed commands.

api-integration-patterns