cleaning-up-branches

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.70). The skill explicitly fetches and inspects remote repository data (git fetch --prune, git branch -r, git cherry, git log, etc. in SKILL.md/WORKFLOW.md Steps 3, 5–7), which ingests untrusted/user-generated third-party content (remote branch names, commit messages) and uses that data to decide and perform destructive actions (deleting branches).