implement

SUSPICIOUS: mostly coherent implementation workflow, but it grants an AI agent meaningful autonomy to modify code, spawn subagents, install packages, and push branches to origin. Risks stem from operational scope and transitive trust, not clear malware or credential theft.