From tdp
Implement plan.md tasks in order by delegating coding and categorized AC review to subagents. Use when plan.md already exists and each task should loop implement -> parallel category review reports -> read review reports until COMPLETE, while the final build task starts from build review first, where n must be replaced with the actual task number, without running final test-plan verification.
npx claudepluginhub bryant-tang/turbo-plugins-claude --plugin tdpThis skill uses the workspace's default tool permissions.
- A requirement already has `plan.md`.
Mandates invoking relevant skills via tools before any response in coding sessions. Covers access, priorities, and adaptations for Claude Code, Copilot CLI, Gemini CLI.
Share bugs, ideas, or general feedback.
plan.md.test-plan.md should not run yet.plan.md or prefer a single-session implementation without subagent review loops.plan.md is identified.n replaced by the actual task number.COMPLETE.plan.md to use. If more than one candidate fits, ask the user instead of guessing.csharp-comment skill and add or update the required XML documentation comments plus any needed single-line or multi-line explanatory comments accordingly.write-plan: Correctness, Security, and Integration; Maintainability, Testability, and Observability; Performance, Resource Usage, and User Experience.COMPLETE, the current task is not COMPLETE.COMPLETE, run another implementation subagent for that same task, then rerun the parallel category reviews, and let each review subagent overwrite its own report file.COMPLETE after 3 attempts, mark it as BLOCKED and stop. Report the blocking findings to the user before proceeding to the next task.test-plan.md verification in this skill.task-n-<category>-review.md, replacing n with the actual task number and <category> with a stable lowercase kebab-case category name such as correctness-security-integration, maintainability-testability-observability, or performance-resource-ux.task-n-build-review.md, replacing n with the actual task number.reviews/ folder inside the same spec folder.plan.md. If ambiguous, ask the user.plan.md and the sibling goal.md.csharp-comment skill.COMPLETE, move to the next task.COMPLETE, feed the blocking findings from those report files back into a new implementation subagent for the same non-build task, then rerun the parallel category review subagents and let them overwrite their own report files. Do not exceed 2 additional implementation attempts after the first review (3 total); if still not COMPLETE, record the task as BLOCKED and stop.task-n-build-review.md.COMPLETE, invoke an implementation subagent limited to fixing the reported build failures, then rerun the build-focused review subagent and let it overwrite the same build review report.COMPLETE, then continue.test-plan.md here.plan.md first instead of silently doing a mega-task.csharp-comment as a required implementation rule rather than an optional polish item.test-plan.md verification.n replaced by the actual task number in the file path and visible title.COMPLETE.task-n-build-review.md report and its verdict is COMPLETE.After all tasks are marked COMPLETE, if end-to-end verification is needed, invoke /tdp:testing-and-proof next to execute test-plan.md.