Captain Claude Code — Testing skill for Claude Code
Captain Claude Code (CCC): a Claude skill that takes a non-technical founder from a product spec to a working product — phased plan, independent verification, plain-language hand-offs.
How to install Captain Claude Code
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open liuyuze1012-code/captain-claude-code and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Captain Claude Code does
Captain Claude Code (CCC): a Claude skill that takes a non-technical founder from a product spec to a working product — phased plan, independent verification, plain-language hand-offs.
Alternatives in Testing
- OpenAgentsControl — AI agent framework for plan-first development workflows with approval-based execution 4.8k ★
- Spec Driven Development — Drive Kandev feature work through spec, plan, independent tasks, implementation, QA, and verification 697 ★
- Maestro Spec — Intent-driven spec precipitation — state a constraint in natural language (加一条规范:禁止用 any / 记录架构约束:服务间走 gRPC / 530 ★
README
Captain Claude Code (CCC) ⚓
**A Claude skill that takes a non-technical founder from a product idea to a working product — in one guided, mostly-autonomous run.**
You describe what you want to build (or hand CCC a design/system spec). CCC plans the build in phases, executes and independently verifies each phase itself, and only stops when a step genuinely needs *your* hands — creating an account, pasting an API key into a `.env` file, clicking a deploy button. Every stop comes with a plain-English why, the exact link, and click-by-click steps.
Born from a real shipped project: a WeChat mini-program built end-to-end by a student founder with no coding background, using this exact workflow.
What it does
- Interactive intake — CCC greets you and asks: "Do you have a design or system spec, or should we create one together?" Every question is answerable by a non-technical person, but each answer lets CCC make a real engineering decision (it asks "will people share things with each other?", not "SQL or NoSQL?").
- Senior-engineer analysis behind the scenes — user scale, per-user data privacy and security, system design, and "is this shippable or just a demo," with online research when decisions depend on current tools or pricing.
- A phased plan ordered by risk — the scariest unproven assumption (payments, a platform API) is tackled first, so surprises come early and cheap.
- An autonomous build loop with honest verification — every phase's work is graded by a separate verifier that never saw the implementation, against written, testable acceptance criteria. Failures are fixed and re-verified automatically; you're only interrupted for product decisions.
- An append-only decisions log — every non-obvious choice, its rejected alternatives, and the why, so nothing gets silently reversed later.
- Progress you can read — regular status blocks: what CCC is working on now, real percent complete, what's next, and whether it needs anything fro
Related Skills
PM Working Backwards Agent
Multi-agent CrewAI pipeline that turns a product problem statement into a research brief, PRFAQ, BRD, and buil
Rpi Plan
RPI Step 3 — Generate comprehensive planning documentation for a feature that received a GO. Produces pm.md (p
Hookwright Form Builder
Claude skill that turns a plain-language form description into a Deftmind Hookwright field spec (JSON) for bui
Intent Planner
Pre-spec steering layer for AI coding agents — capture product intent, invariants, and decision criteria befor
Pre Commit Review
Independent QA review of a working-tree diff before commit. Spawns a read-only reviewer agent to find correctn
Spec Specify
Phase 1 — Creates the detailed technical specification of a feature: functional requirements, non-functional r
Related Agents
CodeHelper
Explains broken or confusing code in plain language and proposes a minimal fix. USE WHEN code isn't working, e
Evidence Collector
Independent verification agent. Takes FINAL-PLAN.md and independently verifies every acceptance criterion was
Verifier Opus
Independent verification of a completed change — does the diff actually do what the plan said, do build/tests/