What problem does it solve? Teams running autonomous delivery loops need a zero-config baseline that decides what to build next, what is ready to promote, and what is structurally broken in the backlog — without injecting opinionated methodology bias. This Skill answers those questions purely from the tracker dependency graph. ## Core Features & Use Cases - Graph-derived ordering: Ranks issues by priority then chainable unlock value (live blocker edges only), with optional PRD-distance composition as a within-band tiebreaker. - Backlog diagnostics: Detects dependency cycles, ghost-project pointers, repeat-park patterns, splittable specs, and chain gaps, with conservative mechanical fixes gated by appetite. - Delivery outputs: Provides promotion-readiness checks, ticket shaping from discovery briefs, standup digests, horizon assignment, build-queue assembly, run-termination policy ladders, crank-up sets, PRDR authoring, and YAGNI judgement. - Use Case: During an unattended L3 drain run, the orchestrator requests build-queue to get a conflict-analysed, wave-structured work order, then calls backlog-diagnostics to demote repeat-parked issues and surface chain gaps before the next wave starts. ## Quick Start Configure the methodology slot to faffter-noon-methodology-thematic and request the pick-ordering output to rank your tracker backlog by unlock value.
