The Agent Skills Directory

[SAFE]: The skill implements a comprehensive risk-management framework (R0-R5) for Git operations. It mandates state inspection (e.g., git status, git branch -vv) and verification for all actions, ensuring that user work is protected.
[COMMAND_EXECUTION]: The skill uses standard Git commands for repository management. High-risk operations, such as force-pushing or remote branch deletion, are gated behind strict confirmation requirements and safety protocols, such as preferring force-with-lease over a bare force-push.
[DATA_EXFILTRATION]: Network operations are strictly limited to standard Git remote interactions (fetch, push, ls-remote) that are necessary for the skill's purpose and require explicit user intent. No unauthorized exfiltration or suspicious network patterns were found.

branch-workflow