spec-first-superpowers

Enforces spec-before-code workflow for AI-driven development. Automatically selects Spec-Kit or OpenSpec mode, triages complexity (quick/standard/thorough), recovers session context, and applies quality gates (G0-G4) at every stage. Use this skill whenever the user says "/super-spec", "spec first", "规范先行", or starts any feature, bugfix, or refactor — especially in projects with .spec-mode, .specify/, or .openspec/ directories. Even if the user doesn't explicitly ask for spec-driven workflow, activate this skill for any non-trivial code change to prevent skipping the design phase. Orchestrates: Spec-Kit/OpenSpec + planning-with-files + ui-ux-pro-max + Superpowers (TDD, code review, verification, debugging).

Safety Notice

This listing is imported from skills.sh public index metadata. Review upstream SKILL.md and repository scripts before running.

Copy this and send it to your AI assistant to learn

Install skill "spec-first-superpowers" with this command: npx skills add zxzvsdcj/spec-first-superpowers/zxzvsdcj-spec-first-superpowers-spec-first-superpowers

Spec-First + Superpowers Orchestrator v3

Stop the AI from jumping straight to code. Every feature, bugfix, and refactor goes through a specification phase first — because unexamined code is expensive code.

Commands

CommandEffect
/super-specFull workflow (auto mode + auto complexity)
/super-spec force-spec-kitForce Spec-Kit mode
/super-spec force-openspecForce OpenSpec mode
/super-spec resetReset mode selection

How It Works

Step 1: Pick a Mode

Check for existing signals, then fall back to heuristics:

SignalMode
.spec-mode file existsUse whatever it says
.specify/ directorySpec-Kit
.openspec/ directoryOpenSpec
Brand new project, < 30 filesSpec-Kit
Everything elseOpenSpec (default)

Save the choice to .spec-mode so future sessions remember it.

For detailed mode workflows, read:

Step 2: Triage Complexity

The AI suggests a level; the user confirms or overrides.

LevelWhenWhat happens
QuickSingle-file bugfix, typo, configSimplified spec (/opsx:ff) → TDD → archive
StandardSingle feature, clear scopeAll phases (Phase 3 only if UI)
ThoroughMulti-module, architecture decisionsAll phases + Agent Teams evaluation

Step 3: Execute the Pipeline

Phase 0 — Session Recovery (automatic) If task_plan.md exists from a previous session, read all planning files, run the 5-Question Reboot Test (Where am I? / Where am I going? / What's the goal? / What did I learn? / What did I do?), then resume from the last checkpoint.

Phase 1 — Specification Write the spec using the selected mode. Quick tasks use fast-forward; standard/thorough use the full flow. The user must explicitly confirm the spec before moving on. Gate G1: User confirmed + spec aligns with constitution.

Phase 2 — Persistent Planning Generate task_plan.md (numbered checklist with test points), findings.md, and progress.md using planning-with-files + writing-plans. Gate G2: Every task has file paths + acceptance criteria + test strategy.

Phase 3 — UI/UX Design (conditional) Triggered only when UI keywords are detected. Invoke ui-ux-pro-max --design-system --persist to generate and persist the design system. Gate G3: Pre-delivery checklist passed + user confirmed design.

Phase 4 — Implementation Execute via one of two strategies (AI recommends, user picks):

  • Subagent-Driven: Fresh subagent per task + two-stage review (spec conformance → code quality)
  • Executing-Plans: Batch execution + checkpoint reviews

TDD throughout. Errors escalate through the 3-Strike protocol → systematic-debugging. Gate G4: All tests pass + review passed + verification evidence written to progress.md.

Phase 5 — Archive finishing-a-development-branch → update all checkboxes → archive spec artifacts → final progress.md entry.

Quality Gates

Each gate is a hard stop — nothing moves forward until all checks pass. If a gate fails, fix the issue and re-evaluate. Full gate criteria: references/quality-gates.md

Anti-Rush Protection

If the user asks to skip the spec phase, politely decline and redirect to /super-spec. The whole point of this skill is preventing premature implementation.

Reference Files

Read these as needed — they contain detailed procedures that would bloat this file:

FileWhen to read
references/quality-gates.mdEvaluating any gate (G0-G4)
references/synergy-patterns.mdUnderstanding cross-tool integration
references/integration-guide.mdSetup, troubleshooting, dependency list
references/spec-kit-workflow.mdRunning the Spec-Kit flow
references/openspec-workflow.mdRunning the OpenSpec flow
assets/constitutions/openspec-constitution.mdOpenSpec constitution template
assets/constitutions/spec-kit-constitution.mdSpec-Kit constitution template

Source Transparency

This detail page is rendered from real SKILL.md content. Trust labels are metadata-based hints, not a safety guarantee.

Related Skills

Related by shared tags or category signals.

General

miniqmt-skills

No summary provided by upstream source.

Repository SourceNeeds Review
Coding

openclaw-version-monitor

监控 OpenClaw GitHub 版本更新,获取最新版本发布说明,翻译成中文, 并推送到 Telegram 和 Feishu。用于:(1) 定时检查版本更新 (2) 推送版本更新通知 (3) 生成中文版发布说明

Archived SourceRecently Updated
Coding

ask-claude

Delegate a task to Claude Code CLI and immediately report the result back in chat. Supports persistent sessions with full context memory. Safe execution: no data exfiltration, no external calls, file operations confined to workspace. Use when the user asks to run Claude, delegate a coding task, continue a previous Claude session, or any task benefiting from Claude Code's tools (file editing, code analysis, bash, etc.).

Archived SourceRecently Updated