What problem does it solve? When an AI agent team stalls, loops infinitely, or recovers work too aggressively, managers need a root cause and a durable product rule rather than a one-off patch. This Skill provides a repeatable forensic procedure that identifies the exact stop point in an issue tree and turns it into an approved, phased plan. ## Core Features & Use Cases - Tree forensics: Walks the linked issue, its blocker chain, recovery siblings, and recent runs to pinpoint the exact issue and state combination that stopped progress, quoting run ids and comment timestamps as evidence. - Invariant-checked rule design: Frames every fix as a general product contract that preserves three invariants: productive work continues, only real blockers stop work, and no infinite loops. - Approval-gated planning: Writes a phased plan into the issue's plan document, opens a request_confirmation interaction, and only decomposes into child issues after board or CTO acceptance. - Use Case: A manager links a stalled issue tree and asks why it stopped. The Skill walks the tree, finds an in_review leaf with no execution participant, surveys recently shipped liveness work, and proposes a contract update to doc/execution-semantics.md with phased implementation issues. ## Quick Start Ask the agent to diagnose why the linked Paperclip issue tree stopped working and produce a root-cause write-up with a proposed product rule before any code changes.
