MageByte Power Skills

Battle-tested Claude Code Superpowers skills — engineering workflows and verification methodologies distilled from real production incidents.

---

Skills

Skill	What it does	When to use
cross-verified-feature-development	7-phase feature workflow with 4 independent cross-verification passes	High-stakes backend features: payments, state machines, concurrency, cross-service changes

---

What is Superpowers?

Superpowers is an MCP plugin for Claude Code that provides general-purpose skills: brainstorming, writing-plans, subagent-driven-development, systematic-debugging, code-reviewer, and more.

The skills in this repo are domain-specific orchestrators — they call Superpowers' general skills at the right phase with the right context, so you get the full power of the ecosystem without wiring it up yourself.

You (user prompt)
    │
    ▼
cross-verified-feature-development  ← this repo
    │
    ├── Phase 1   →  superpowers:brainstorming
    ├── Phase 2   →  superpowers:writing-plans
    ├── Phase 3   →  superpowers:subagent-driven-development
    ├── Phase 4.1 →  superpowers:systematic-debugging
    ├── Phase 4.2 →  superpowers:code-reviewer  (cold-context, no design docs)
    ├── Phase 4.3–4.5 → dispatches agents with domain-specific prompt templates
    └── Phase 5   →  superpowers:writing-plans + subagent-driven-development

Without Superpowers, the skill still works — each phase has a documented fallback using Claude Code's built-in tools.

---

Prerequisites

1. Claude Code

npm install -g @anthropic-ai/claude-code

2. Superpowers plugin (recommended)

claude mcp add --transport http superpowers https://superpowers.anthropic.com/mcp

First-time use requires an Anthropic account. Fallback mode is available without Superpowers.

---

Installation

git clone https://github.com/MageByte-Zero/magebyte-power.git
export SKILLS_REPO="$PWD/magebyte-power"

mkdir -p ~/.claude/skills

ln -sf "$SKILLS_REPO/skills/cross-verified-feature-development" \
       ~/.claude/skills/cross-verified-feature-development

Verify:

ls ~/.claude/skills/cross-verified-feature-development/SKILL.md

---

cross-verified-feature-development

A rigorous 7-phase feature development methodology that inserts 4 independent verification passes between implementation and merge — catching the concurrency, idempotency, and cross-service contract bugs that self-review and standard code review systematically miss.

The core problem this solves

Single-perspective review has structural blind spots. Even experienced developers will:

• Default to their own design assumptions when reading their own code
• Assume "should be idempotent" operations actually are
• Miss concurrent race windows because their mental model is sequential
• Forget that other services consume their shared tables and MQ topics

Independent perspectives produce independent signals. Running N reviewers who don't share context produces a bug set that's close to a union, not a duplicate. This workflow operationalizes that insight:

Design → Implement → Multi-round cross-verification → Fix → Carefully simplify → Sync docs

How Superpowers powers each phase

Phase	Superpowers skill	What it adds
1. Requirements & Design	superpowers:brainstorming	Structures vague requirements into a complete spec with invariants, failure modes, and risk table
2. Implementation Plan	superpowers:writing-plans	Breaks spec into independently verifiable tasks with file+line targets, TDD structure, rollout strategy
3. Implementation	superpowers:subagent-driven-development	Each task runs in a fresh subagent — no accumulated bias, independent spec compliance + quality review per task
4.1 Self-review	superpowers:systematic-debugging	Structured "assume there's a bug, find it" scan against the full feature
4.2 Cold-context review ⭐	superpowers:code-reviewer	Reviewer gets no design docs — finds design-level blind spots the author can't see
4.3–4.5 (parallel)	Custom agent prompts	Behavior-preservation diff, cross-repo impact scan, business invariant matrix
5. Fix iteration	superpowers:writing-plans + subagent-driven-development	Same rigor for fixes as for original implementation

Phase 4.2 (Cold-Context Review) is the highest-value step. Typical yield: 5–15 Critical/High bugs per review that would have reached production with standard code review.

When to use

Does the feature involve any of the following?
│
├── 💰 Financial transactions, payments, refunds, settlements?      → YES → Use this workflow
├── 🔄 Order / inventory state machines with status transitions?    → YES → Use this workflow
├── 🔒 Distributed locks, concurrency control, idempotent retry?    → YES → Use this workflow
├── 🔗 Cross-service MQ/RPC contracts or shared proto/model change? → YES → Use this workflow
├── 🗄️  Online schema migration or dual-write strategy?             → YES → Use this workflow
└── ⏱️  Estimated effort ≥ 3 person-days with high cost-of-failure? → YES → Use this workflow

None of the above?  →  Standard workflow is fine ✓

One-line heuristic: If the worst possible bug involves money loss / data corruption / stuck orders / privilege escalation, use this workflow.

The 7 phases

① Requirements & Design    → superpowers:brainstorming
①.5 Architecture Review    → ADR (required for high-risk features)
② Implementation Plan      → superpowers:writing-plans
③ Implementation           → superpowers:subagent-driven-development
④ 🔥 Cross-Verification    ← 4 independent passes (the core innovation)
⑤ Fix Iteration            → writing-plans + subagent-driven-development
⑥ Careful Simplification   → skeptical optimization with anti-pattern checklist
⑦ Doc Sync                 → backfill evolution log + notify downstream

The 4 verification passes

Pass	Perspective	Typical findings
4.1 Systematic Debugging	Self as bug hunter	1–3 pre-existing bugs
4.2 Cold-Context Review ⭐	Reviewer with no design docs	5–15 concurrency / idempotency bugs
4.3 Behavior-Preservation Diff	Master vs feature side-effects	2–5 semantic regressions
4.4 Cross-Repo Impact Scan	Other repos that may need changes	0–3 external impact points
4.5 Business Invariant Matrix	Hard constraints (money/state/inventory)	0–2 invariant violations

Passes 4.3–4.5 can be dispatched in parallel via subagents.

How to trigger

/cross-verified-workflow <feature description>

Or describe a high-stakes feature — the skill proactively suggests itself when it detects relevant patterns.

Bundled reference files

File	When to read
references/cross-verification-techniques.md	Before Phase 4 — agent prompt templates for all 4 passes
references/anti-patterns.md	Before Phase 6 — 12 high-frequency trap patterns
references/doc-sync-playbook.md	Before Phase 7 — structured doc backfill
references/case-studies.md	Optional — real-world bug museum from e-commerce order domain

Cost vs. benefit

Metric	Standard workflow	This workflow
Extra time	—	+40–50%
Critical bug detection rate	~40%	~95%
Phase 4.2 typical yield	0	5–15 High/Critical bugs

---

Contributing

Skills in this repo are distilled from real engineering experience. The best contributions are workflows proven in production.

1. Create skills/<skill-name>/SKILL.md with required frontmatter:

   ---
   name: <skill-name>
   description: <when to trigger — be specific about contexts and keywords>
   ---

2. Add supporting files in skills/<skill-name>/references/
3. Test it on real tasks before submitting
4. Open a PR describing the domain and what problem it solves

---

📖 Read the full story (Chinese): 我把 4 年踩坑经验「蒸馏」成 Claude Code Skill 开源了 — how production incidents shaped every phase of this workflow.

---

Maintainer: MageByte-Zero · MIT License