What problem does it solve? Manually coordinating specification, planning, implementation, and review of a GitHub issue across many tools and chat prompts is error-prone and loses context. This Skill automates the entire lifecycle of a single GitHub issue using the gh CLI, keeping every decision and approval on the issue or PR thread instead of the console. ## Core Features & Use Cases - End-to-end issue pipeline: Sequences the speckit phases (specify, clarify, plan, tasks, implement, test) from claiming an issue through opening a draft PR, watching CI, and handling reviewer comments until merge. - Human-in-the-loop approval gates: Posts every artefact (spec, plan, tasks) as a marked comment on the issue and polls for owner approval before proceeding, never auto-answering clarification questions. - Strict communication protocol: Enforces owner-login filtering, watermark-based polling, and re-fetch-before-post rules so no owner feedback is ever missed. - Use Case: Tell the agent "implement issue #42 with speckit" and it claims the issue, drafts and gets approval for the spec, plan, and tasks, opens a draft PR, implements and tests the code, watches CI, and responds to reviewer comments until the owner directs a merge. ## Quick Start Ask the agent to implement a specific GitHub issue end-to-end with speckit, for example by saying "implement issue #123 with speckit" or pasting the issue URL.
