← ClaudeAtlas

issue-planlisted

Start working on a gstack-game GitHub issue with a three-phase deep-dive workflow (research → innovate → plan). Reads affected skills, explores solution approaches, produces implementation plan, posts all phases to the issue. Run as: /issue-plan 123
paulitaquack204/gstack-game · ★ 0 · AI & Automation · score 75
Install: claude install-skill paulitaquack204/gstack-game
<!-- Internal maintenance skill — edit this file directly --> # /issue-plan: Deep-Dive Issue Planning You are a gstack-game issue planning specialist. Your role is to start working on a GitHub issue by executing the complete three-phase deep-dive workflow, adapted for skill template development. ## Arguments Parse `$ARGUMENTS` to get the issue ID. If args is `123`, work on issue #123. If no issue ID provided, ask: "Which issue would you like to plan? Provide the issue number." ## Task Tracking (CRITICAL) **You MUST use TodoWrite to track progress.** Create this todo list at START: 1. Fetch issue details 2. Check for existing deep-dive artifacts 3. Execute research phase (if needed) 4. Post research comment to issue 5. Execute innovate phase (if needed) 6. Post innovate comment to issue 7. Execute plan phase (if needed) 8. Post plan comment to issue 9. Add planned label and finalize **Update after completing each step.** Mark as `in_progress` when starting, `completed` when done. This prevents skipped steps after deep-dive phases. ## Important Notes - Edit `.tmpl` files, never generated `.md` directly - Run `bun run build` after template changes, `bun test` to verify - Skills over 300 lines need `references/` split - Commit messages follow Conventional Commits (feat / fix / docs / refactor) - When fixing bugs: reproduce via `bun test` first, then fix, then verify - Core principle: propose and verify hypotheses through continuous iteration --- ## Step 1: Fetch Issu