autoplan

SUSPICIOUS. The core review purpose matches reading repo state and sibling review skills, and the gstack-local tooling appears broadly coherent. But the skill’s footprint is unusually broad for a review pipeline: it executes many helper binaries, can edit project guidance, commit changes, invoke other skills transitively, and auto-decide actions in spawned sessions. This is not confirmed malware, but it is a medium-risk skill with autonomy and trust-expansion concerns.