The Agent Skills Directory

[SAFE]: The skill establishes a standardized naming convention for Git branches and focuses on workflow hygiene without introducing malicious commands or patterns.
[DATA_EXFILTRATION]: The instructions explicitly prohibit including secrets, customer names, credentials, or private prompt text in branch names, directly addressing and mitigating data leakage risks.
[COMMAND_EXECUTION]: The skill uses standard Git commands (e.g., git status, git branch, git remote) for context gathering and recommends git check-ref-format for validating branch names, which is a defensive practice to prevent malformed or malicious branch names.
[INDIRECT_PROMPT_INJECTION]: While the skill interacts with Git metadata which could be externally influenced, it provides clear boundary markers and validation steps to ensure branch names are safe and conform to expected formats, minimizing the surface for indirect injection.

branch-strategy