The Agent Skills Directory

[SAFE]: The skill defines procedures for standard Git operations like merging, rebasing, and fast-forwarding. It does not include commands for unauthorized network access or credential access.
[SAFE]: The skill identifies and mitigates potential Indirect Prompt Injection vectors. 1. Ingestion points: Processes repository metadata (commit messages, hooks, issue text) and upstream source code. 2. Boundary markers: Instructions explicitly state that repository data should not be treated as authoritative unless established independently. 3. Capability inventory: Operations are limited to Git commands and defined verification steps. 4. Sanitization: Verification of graph shape, diffs, and test results is required post-integration.
[SAFE]: No obfuscation, persistence mechanisms, or privilege escalation patterns were detected in the instructions or reference files.

integrate-branches