The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the subprocess module to run git commands (e.g., git diff, git commit, git push) locally. These calls are implemented using argument lists rather than shell strings, following security best practices to prevent command injection.
[DATA_EXFILTRATION]: The skill retrieves API tokens from the environment, git configuration, and local configuration files. These credentials are used exclusively to communicate with the official APIs of GitHub and GitLab to automate project workflows. No data is sent to unauthorized third-party domains.
[INDIRECT_PROMPT_INJECTION]: The skill analyzes local repository data, such as code diffs and commit messages, to generate summaries. While this processes external content that could theoretically contain malicious instructions, the extraction logic is based on deterministic regular expressions rather than LLM-based interpretation within the script logic, minimizing the risk of accidental obedience to embedded instructions.

create-pr