The Agent Skills Directory

[SAFE]: The skill is designed to improve Git hygiene through structured workflows. It adheres to standard developer practices and does not exhibit any malicious behaviors such as credential harvesting, obfuscation, or unauthorized exfiltration.
[COMMAND_EXECUTION]: The skill defines several shell-based workflows using Git commands. These are used for repository maintenance, including atomic committing, interactive rebasing (utilizing the GIT_SEQUENCE_EDITOR environment variable for automation), and history searching. All commands are standard for development and restricted to the repository context.
[PROMPT_INJECTION]: The skill uses authoritative language (e.g., 'CRITICAL', 'NON-NEGOTIABLE', 'HARD RULE') to enforce Git standards. These instructions are part of the agent's task-specific logic and do not attempt to override or bypass platform-level safety constraints or system prompts.

git-master