OPC — One Person Company

A full team in a single Claude Code skill. You're the CEO — OPC is everyone else.

16 specialist agents (PM, Designer, Security, Devil's Advocate, and more) that build, review, and evaluate your code through a digraph-based pipeline with code-enforced quality gates.

English | 中文

English

What you can build

From a single one-line brief, OPC's build-verify flow ships a complete, production-quality product — design tokens injected and visual quality validated by the design-intelligence extension. Every screenshot below is a real, clickable site (not a mockup), built by the same 16-agent pipeline. Same system, six completely different design languages.

▶ Browse the live lookbook →

Pulse · SaaS analytics dashboard _{Dark UI · data-viz}	Studio Kura · Creative agency portfolio _{Brutalist · editorial · motion}
Cumulus · Project-management dashboard _{Claymorphism · pastel}	Conduit Docs · Developer documentation _{Docs · code · developer}
Maison Terre · DTC skincare e-commerce _{Brand · editorial}	Yuzuki 柚月 · Local dining platform _{Booking · bilingual · mobile}

Each card in the lookbook flips on hover to reveal the exact prompt that built it — e.g. Pulse came from one line: /opc build a production-quality SaaS analytics dashboard called "Pulse"...

OPC isn't only for UIs. The same zero-trust pipeline also handles:

Whole products, unattended — /opc loop build features F1–F4 from PLAN.md decomposes the work, schedules a durable cron, and runs 10+ hours without you.
Code review — /opc review the auth changes dispatches 2–5 independent role agents (security, backend, a11y, devil's advocate) in parallel and computes a mechanical verdict.
Full-stack features — /opc implement user auth with email/password runs build → independent review → test-design → test-execute → gate.
Pre-release audits — /opc verify before release runs acceptance + audit + e2e gates before you ship.

How It Works

One principle: the agent that does the work never evaluates it.

Task → Flow Selection → Node Execution → Gate Verdict → Route Next
                              ↑                              ↓
                              └──────── ITERATE/FAIL ────────┘

Task inference — reads your request, picks a flow template (quick, review, build-verify, full-stack, pre-release), and enters at the right node.
Typed nodes — each node has a type (discussion, build, review, execute, gate) with specific protocols. Build nodes produce commits. Review nodes dispatch parallel subagents. Gate nodes compute verdicts from code, not LLM judgment.
Mechanical gates — verdicts are computed by opc-harness synthesize: any red = FAIL, any yellow = ITERATE, all green = PASS. No LLM gets to decide if a finding is "important enough."
Cycle limits — max 3 loops per edge, 5 re-entries per node, 20-30 total steps depending on flow. Oscillation detection catches A↔B loops.

Quality Architecture

The system is built on a zero-trust axiom: every critical output must have an independent verification path. Four layers:

L0 — Zero Trust: Decision axiom — not code, not prompt. Every critical output needs an independent verification path.
L1 — Shape Single Agent: Intervene during token generation — persona setting, anti-pattern tables, mandatory output structure, scope anchoring, quality gates.
L2 — Design Agent Flow: Multi-agent coordination — separation of concerns, flow topology (parallel review / sequential build), context isolation (file-based handoff, fresh agents, no session reuse).
L3 — Deterministic Enforcement: The only layer that doesn't need LLM compliance — mechanical ops (severity counting, verdict rules, oscillation diff) and hardened verification (file-finding evidence checks, hedging scans, ref validation).

Quick Start

Install

npm install -g @touchskyer/opc

Skill files are automatically copied to ~/.claude/skills/opc/.

Manual install (no npm)

git clone https://github.com/iamtouchskyer/opc.git
cp -r opc ~/.claude/skills/opc

Use it

# Review — dispatches 2-5 role agents in parallel
/opc review the auth changes

# Build — implements + independent review + gate
/opc implement user authentication with email/password

# Autonomous loop — decomposes, schedules cron, runs unattended
/opc loop build features F1-F4 from PLAN.md

# Interactive mode — asks clarifying questions first
/opc -i redesign the onboarding flow

# Explicit roles
/opc security devil-advocate

# Flow control
/opc skip          # skip current node
/opc pass          # force-pass gate
/opc stop          # terminate, preserve state
/opc goto build    # jump to node

Autonomous Loop

/opc loop build the math tutoring app features F1-F4

What happens:

Runbook lookup — opc-harness runbook match "<task>" checks --dir flag → OPC_RUNBOOKS_DIR → ~/.opc/runbooks/ for a matching recipe. If one hits, its units / flow / tier become the plan; otherwise fall through to step 2. Disable per-invocation with OPC_DISABLE_RUNBOOKS=1. See docs/runbooks.md and examples/runbooks/add-feature.md.
Decompose (runbook miss only) — breaks task into atomic units (spec, implement, review, fix, e2e)
Definition of done — establishes verify/eval criteria per unit before any work starts
Schedule — durable cron (survives process restart) fires every 10 min
Execute — each tick runs one unit through the appropriate OPC flow
Guard — opc-harness enforces: git commit required, ≥2 independent reviewers, no plan tampering, no state forgery, artifact freshness, tick limits
Terminate — auto-stops when plan complete, tick limit hit, or wall-clock deadline reached

For well-scoped tasks, the system runs 10+ hours continuously without intervention.

Guardrails (code-enforced, not prompt-level)

Guard	Enforcement
Write nonce	Random SHA256 at init; state written by harness only
Atomic writes	write → rename (POSIX atomic); no truncated JSON on crash
Plan integrity	SHA256 hash at init; verified every tick
Review independence	≥2 eval files, identical content rejected, line overlap warned
Git commit required	HEAD must change for implement/fix units
Screenshot required	UI units must produce .png/.jpg artifact
Tick limits	maxTotalTicks (units×3) + 24h wall-clock deadline
Oscillation detection	A↔B pattern over 4-6 ticks = warning/hard stop
Concurrent tick mutex	in_progress status blocks overlapping cron fires
JSON crash recovery	try/catch on all JSON.parse; structured errors, not crashes
External validators	Pre-commit hooks, test suites detected at init and leveraged

Flow Templates

Template	Nodes	When
quick	build → review → gate	"quick fix", "small change", one-liner, ≤3 files, non-UI, no logic branches
review	code-review → gate	PR review, audit, "find problems"
build-verify	brief → build → code-review → test-design → test-execute; ITERATE → hotfix → test-execute; PASS → gate	"implement X", "fix bug Y"
full-stack	discuss → build → review → test → acceptance → audit → e2e → gates	Complex/vague requests
pre-release	acceptance → audit → e2e → gates	"verify before release"

quick vs build-verify — quick drops the structured brief node and the test-design/test-execute split, halving nodes and evaluator dispatches. It keeps the core quality gate (independent review + gate verdict). Use it only for true one-liners / single-file, non-UI, no-logic-branch changes. Anything with logic branches, multiple files, or UI → use build-verify, where test-design's enforced L1–L5 coverage check earns its keep.

The brief node (entry of build-verify) produces a structured build brief — resolved design tokens, file plan, component inventory, constraints — that the build node follows. It measurably lifts first-pass build fidelity, but the brief describes intended behavior: downstream test-design must test the actual built code, not the brief's ideal, or it will assert behavior the build never implemented.

Extensions

OPC has a capability-routed extension surface. Extensions live in ~/.claude/skills/opc-extension/<name>/ — each with ext.json (capability declarations) + hook.mjs exporting any of promptAppend / verdictAppend / executeRun / artifactEmit hooks. No fork, no rebuild. Hooks are sandboxed via per-extension timeouts + circuit breakers, so a broken third-party extension can't take down the harness.

ext.json.version is recorded in flow-state.json → extensionVersions during init, so protocol reports can distinguish "extension loaded" from "unknown version". Executable hook metadata from hook.mjs still wins for runtime capability behavior.

The companion repo opc-extensions ships 4 extensions: design-intelligence (theme injection + design coverage + VLM visual eval), git-changeset-review, memex-recall, and session-logex.

Full authoring guide: docs/extension-authoring.md — zero-OPC-context quickstart + reference, plus a starter template at examples/extensions/_starter/.

Built-in Roles

Product:     pm, designer
User Lens:   new-user, active-user, churned-user
Engineering: frontend, backend, devops, architect, engineer
Quality:     security, tester, compliance, a11y
Specialist:  planner, user-simulator, devil-advocate

Devil's Advocate (the 10th person) is auto-included when consensus is near-unanimous or decisions are irreversible. Comes with an automated verification script that checks its own findings.

Custom Roles

Add a .md file to roles/:

---
tags: [review, build]
---
# Role Name

## Identity
One sentence: who you are and what you care about.

## Expertise
- **Area** — what you know about it

## When to Include
- Condition that triggers this role

Available immediately, no configuration needed.

Testing

bash test/run-all.sh

100+ test files covering init-loop, complete-tick, next-tick, review independence, JSON crash recovery, compound defense, scope registry, criteria lint, pipeline E2E lint, D2 calibration, release packaging, and orchestrator-level E2E flow tests.

Requirements

Claude Code (CLI, desktop app, or IDE extension)
Node.js >= 18
Core runtime has no npm dependencies, no MCP server, no build step
Optional: jq for opc install-hooks context-compaction hooks

Works better with memex (optional)

OPC works standalone — pair it with memex for cross-session memory. Memex remembers which roles were useful, which findings were false positives, and your project-specific context.

npm install -g @touchskyer/memex

What's New

See CHANGELOG.md for version history.

Community

Using OPC? Share your setup in Discussions → Show and tell. Questions go in Q&A. Feature ideas in Ideas.

License

MIT

中文

你能用它造什么

从一句话需求出发，OPC 的 build-verify 流程就能交付一个完整、生产级的产品——设计令牌（design token）由 design-intelligence 扩展注入、视觉质量由它验证。下面每一张截图都是真实可点击的站点（不是设计稿），由同一套 16-agent 流水线构建。同一套系统，六种完全不同的设计语言。

▶ 浏览在线作品集 →

Pulse · SaaS 分析仪表盘 _{暗色 UI · 数据可视化}	Studio Kura · 创意工作室作品集 _{粗野主义 · 编辑风 · 动效}
Cumulus · 项目管理仪表盘 _{黏土拟态 · 柔和色彩}	Conduit Docs · 开发者文档站 _{文档 · 代码 · 开发者}
Maison Terre · 护肤品 DTC 电商 _{品牌 · 编辑风}	Yuzuki 柚月 · 本地餐饮平台 _{预订 · 双语 · 移动端}

作品集里每张卡片悬停时会翻面，露出生成它的原始 prompt——比如 Pulse 就来自一句话：/opc build a production-quality SaaS analytics dashboard called "Pulse"...

OPC 不只是用来做 UI。 同一套零信任流水线还能：

无人值守造完整产品 —— /opc loop build features F1–F4 from PLAN.md 拆解任务、调度持久 cron，连续运行 10+ 小时无需干预。
代码审查 —— /opc review the auth changes 并行派发 2–5 个独立角色 agent（security、backend、a11y、devil's advocate），用代码计算裁决。
全栈功能 —— /opc implement user auth with email/password 跑 build → 独立 review → test-design → test-execute → gate。
发布前审计 —— /opc verify before release 在上线前跑 acceptance + audit + e2e 门禁。

工作原理

一条原则：做事的 agent 永远不评判自己的工作。

Task → Flow Selection → Node Execution → Gate Verdict → Route Next
                              ↑                              ↓
                              └──────── ITERATE/FAIL ────────┘

任务推断 —— 读取你的请求，选一个 flow 模板（quick、review、build-verify、full-stack、pre-release），从正确的节点进入。
类型化节点 —— 每个节点有类型（discussion、build、review、execute、gate）和特定协议。build 节点产出 commit；review 节点派发并行 subagent；gate 节点用代码而非 LLM 判断计算裁决。
机械门禁 —— 裁决由 opc-harness synthesize 计算：任意 red = FAIL，任意 yellow = ITERATE，全 green = PASS。没有任何 LLM 能决定某个发现"是否足够重要"。
循环上限 —— 每条边最多 3 次循环，每个节点最多 5 次重入，按 flow 总步数 20-30。震荡检测捕获 A↔B 循环。

质量架构

系统建立在一条零信任公理上：每个关键输出都必须有独立的验证路径。 四层：

L0 —— 零信任：决策公理——不是代码，不是 prompt。每个关键输出都需要独立的验证路径。
L1 —— 塑造单个 agent：在 token 生成过程中干预——人格设定、反模式表、强制输出结构、范围锚定、质量门禁。
L2 —— 设计 agent 流：多 agent 协同——关注点分离、流拓扑（并行 review / 串行 build）、上下文隔离（基于文件的交接、全新 agent、不复用 session）。
L3 —— 确定性强制：唯一不需要 LLM 配合的一层——机械操作（严重度计数、裁决规则、震荡 diff）和加固验证（找文件证据检查、模糊措辞扫描、引用校验）。

快速开始

安装

npm install -g @touchskyer/opc

Skill 文件会自动复制到 ~/.claude/skills/opc/。

手动安装（不用 npm）

git clone https://github.com/iamtouchskyer/opc.git
cp -r opc ~/.claude/skills/opc

使用

# 审查 —— 并行派发 2-5 个角色 agent
/opc review the auth changes

# 构建 —— 实现 + 独立 review + gate
/opc implement user authentication with email/password

# 自主循环 —— 拆解、调度 cron、无人值守运行
/opc loop build features F1-F4 from PLAN.md

# 交互模式 —— 先问清楚需求
/opc -i redesign the onboarding flow

# 显式指定角色
/opc security devil-advocate

# 流程控制
/opc skip          # 跳过当前节点
/opc pass          # 强制通过 gate
/opc stop          # 终止，保留状态
/opc goto build    # 跳转到指定节点

自主循环

/opc loop build the math tutoring app features F1-F4

会发生什么：

Runbook 查找 —— opc-harness runbook match "<task>" 按 --dir flag → OPC_RUNBOOKS_DIR → ~/.opc/runbooks/ 顺序查找匹配的配方。命中则用它的 units / flow / tier 作为 plan；否则进入第 2 步。可用 OPC_DISABLE_RUNBOOKS=1 按次禁用。见 docs/runbooks.md 和 examples/runbooks/add-feature.md。
拆解（仅 runbook miss）—— 把任务拆成原子单元（spec、implement、review、fix、e2e）
完成定义 —— 任何工作开始前，为每个单元建立 verify/eval 标准
调度 —— 持久 cron（进程重启后存活）每 10 分钟触发
执行 —— 每个 tick 跑一个单元，走对应的 OPC flow
守卫 —— opc-harness 强制：必须 git commit、≥2 个独立审查者、不可篡改 plan、不可伪造 state、artifact 新鲜度、tick 上限
终止 —— plan 完成、tick 上限、或墙钟截止时自动停止

对于范围明确的任务，系统可连续运行 10+ 小时无需干预。

守卫（代码强制，非 prompt 层）

守卫	强制方式
Write nonce	init 时生成随机 SHA256；state 只能由 harness 写入
原子写入	write → rename（POSIX 原子）；崩溃时不留半截 JSON
Plan 完整性	init 时算 SHA256 哈希；每个 tick 校验
审查独立性	≥2 个 eval 文件，内容雷同则拒绝，行重叠则告警
必须 git commit	implement/fix 单元的 HEAD 必须变化
必须截图	UI 单元必须产出 .png/.jpg artifact
Tick 上限	maxTotalTicks（units×3）+ 24h 墙钟截止
震荡检测	4-6 个 tick 内出现 A↔B 模式 = 告警/硬停
并发 tick 互斥	in_progress 状态阻止重叠的 cron 触发
JSON 崩溃恢复	所有 JSON.parse 加 try/catch；结构化报错而非崩溃
外部校验器	init 时检测并利用 pre-commit hook、测试套件

流模板

模板	节点	何时用
quick	build → review → gate	"快速修复"、"小改动"、一行改动、≤3 文件，非 UI、无逻辑分支
review	code-review → gate	PR 审查、审计、"找问题"
build-verify	brief → build → code-review → test-design → test-execute；ITERATE → hotfix → test-execute；PASS → gate	"实现 X"、"修复 bug Y"
full-stack	discuss → build → review → test → acceptance → audit → e2e → gates	复杂/模糊的请求
pre-release	acceptance → audit → e2e → gates	"发布前验证"

quick vs build-verify —— quick 砍掉结构化的 brief 节点和 test-design/test-execute 拆分，节点数和评估器派发减半。它保留了核心质量门禁（独立 review + gate 裁决）。只在真正的一行改动 / 单文件、非 UI、无逻辑分支时用。任何带逻辑分支、多文件或 UI 的 → 用 build-verify，那里 test-design 强制的 L1–L5 覆盖检查才物有所值。

brief 节点（build-verify 的入口）产出一份结构化构建简报——解析后的设计令牌、文件计划、组件清单、约束——供 build 节点遵循。它能可量化地提升首次构建保真度，但简报描述的是预期行为：下游 test-design 必须测实际构建出来的代码，而非简报里的理想，否则它会断言构建从未实现的行为。

扩展

OPC 有一个能力路由（capability-routed）的扩展面。扩展位于 ~/.claude/skills/opc-extension/<name>/——每个带一个 ext.json（能力声明）+ 一个 hook.mjs，导出 promptAppend / verdictAppend / executeRun / artifactEmit 中的任意 hook。不用 fork、不用重新构建。Hook 通过每扩展超时 + 熔断器沙箱化，所以一个坏掉的第三方扩展无法拖垮 harness。

ext.json.version 在 init 时记录到 flow-state.json → extensionVersions，所以协议报告能区分"扩展已加载"和"未知版本"。运行时能力行为仍以 hook.mjs 的可执行 hook 元数据为准。

配套仓库 opc-extensions 提供 4 个扩展：design-intelligence（主题注入 + 设计覆盖 + VLM 视觉评估）、git-changeset-review、memex-recall、session-logex。

完整编写指南：docs/extension-authoring.md——零 OPC 上下文的快速上手 + 参考，外加 examples/extensions/_starter/ 里的起步模板。

内置角色

Product:     pm, designer
User Lens:   new-user, active-user, churned-user
Engineering: frontend, backend, devops, architect, engineer
Quality:     security, tester, compliance, a11y
Specialist:  planner, user-simulator, devil-advocate

Devil's Advocate（第 10 个人）在共识近乎一致、或决策不可逆时自动加入。配有一个自动验证脚本，检查它自己的发现。

自定义角色

在 roles/ 里加一个 .md 文件：

---
tags: [review, build]
---
# Role Name

## Identity
One sentence: who you are and what you care about.

## Expertise
- **Area** — what you know about it

## When to Include
- Condition that triggers this role

即刻可用，无需配置。

测试

bash test/run-all.sh

100+ 测试文件，覆盖 init-loop、complete-tick、next-tick、审查独立性、JSON 崩溃恢复、复合防御、scope registry、criteria lint、pipeline E2E lint、D2 校准、发布打包，以及编排器层级的 E2E 流程测试。

环境要求

Claude Code（CLI、桌面应用或 IDE 扩展）
Node.js >= 18
核心运行时无 npm 依赖、无 MCP server、无构建步骤
可选：jq，用于 opc install-hooks 的上下文压缩 hook

搭配 memex 更好（可选）

OPC 可独立使用——搭配 memex 获得跨会话记忆。Memex 记住哪些角色有用、哪些发现是误报，以及你项目特定的上下文。

npm install -g @touchskyer/memex

更新日志

版本历史见 CHANGELOG.md。

社区

在用 OPC？到 Discussions → Show and tell 分享你的玩法。问题发 Q&A，功能想法发 Ideas。

许可

MIT

Name		Name	Last commit message	Last commit date
Latest commit History 309 Commits
bin		bin
docs		docs
examples		examples
pipeline		pipeline
roles		roles
scripts		scripts
test		test
.gitignore		.gitignore
CHANGELOG.md		CHANGELOG.md
CONTRACTS.md		CONTRACTS.md
CONTRIBUTING.md		CONTRIBUTING.md
INTEGRATION.md		INTEGRATION.md
LICENSE		LICENSE
README.md		README.md
SKILL.md		SKILL.md
package.json		package.json
replay.md		replay.md

Folders and files

Latest commit

History

Repository files navigation

OPC — One Person Company

English

What you can build

How It Works

Quality Architecture

Quick Start

Install

Manual install (no npm)

Use it

Autonomous Loop

Guardrails (code-enforced, not prompt-level)

Flow Templates

Extensions

Built-in Roles

Custom Roles

Testing

Requirements

Works better with memex (optional)

What's New

Community

License

中文

你能用它造什么

工作原理

质量架构

快速开始

安装

手动安装（不用 npm）

使用

自主循环

守卫（代码强制，非 prompt 层）

流模板

扩展

内置角色

自定义角色

测试

环境要求

搭配 memex 更好（可选）

更新日志

社区

许可

About

Topics

Resources

License

Contributing

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

Packages