github-cleanup

[Skill Scanner] Backtick command substitution detected This skill is coherent with its stated purpose and uses only gh CLI and local filesystem operations to audit and clean GitHub repositories. It legitimately requires high privileges to perform destructive actions; those privileges are proportionate but represent substantial operational risk. There is no evidence of malicious intent (no obfuscation, no external credential exfiltration, no download-execute patterns). Primary concern: destructive capabilities (repo/secret deletion, disabling security features) and heuristic-based detections that could lead to accidental loss if run without strict user confirmations and dry-runs. Recommend treatment as high-risk operational tool requiring careful human-in-the-loop controls. LLM verification: The selected report presents a coherent, governance-focused GitHub cleanup workflow with appropriate safeguards (user approvals, scoped permissions, and explicit prerequisites). While it acknowledges high-risk actions and documents some potential edge cases, it could be strengthened with explicit per-action confirmations, clearer rollback/audit mechanisms, and stronger RBAC considerations. No malicious behavior is evident in the executable flow; the primary risk is human operator error and misco

This is a safety-focused operational guide that consolidates discovery, verification, approval, execution, and verification steps for destructive GitHub housekeeping. It is not malicious code, but it contains explicit examples of irreversible operations that pose a significant operational risk if copy-pasted into automation or run with improperly scoped/elevated tokens. The main findings: (1) follow the audit-first and approval steps strictly; (2) avoid using --yes or auto-delete loops in unattended scripts; (3) limit tokens to least privilege and log+verify every destructive action. Treat this document as high-impact operational guidance rather than a software package — use with strict manual controls and peer review.