openclaw-pr-batch-sweep

SUSPICIOUS: the skill is coherent with its stated maintainer purpose, but it grants an AI agent high-impact GitHub mutation powers and depends on multiple external tools/skills with unspecified provenance. The main risk is autonomous repository actions plus transitive trust, not confirmed malware.

The code fragment outlines a guarded PR workflow with explicit safeguards but also introduces a meaningful security risk through the possibility of unsigned pushes. The combination of environment-controlled push modes and head reconstruction/overlay logic warrants careful review of the actual implementation, access controls, and validation guarantees. No concrete code is shown, so there is no definitive malware evidence, but the potential for supply-chain abuse via unsigned pushes is non-trivial and should be mitigated.