What problem does it solve? Implementing a large spec-kit task list in one context leads to context bloat, inconsistent commits, and unverified tests. This Skill orchestrates the implementation phase by dispatching each chunk of tasks.md to a fresh subagent, enforcing local test-pass evidence, and producing an auditable final report. ## Core Features & Use Cases - Chunked implementation loop: Splits tasks.md along phase headings and dispatches each chunk to a clean-context subagent with a self-contained brief and response contract. - Mandatory local-pass evidence: Requires every new or modified test to be observed passing locally, with the run command, timestamp, and verbatim pass line recorded. - Review and final report: Runs a full-diff review after all chunks, fixes high-severity findings, and writes a structured opsmill-implement-report.md with a machine-readable STATUS line. - Use Case: After running the speckit prep phase on a feature spec, invoke this Skill to implement all tasks chunk-by-chunk, then receive a ledger of commits, test evidence, and review findings without manually supervising each step. ## Quick Start Run the speckit-opsmill-implement skill on my most recent spec directory under specs/ to implement all tasks and produce the final report.
