Skip to content
Back to search

production-grade

SkillWorks with Claude CodeGitHub

14 AI agents, one install, idea to production: architecture docs, API contracts, tested backend/frontend, security audit, CI/CD pipelines, and documentation — for building, hardening, shipping, and documenting real systems.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/nagisanzenin--production-grade/badge.svg)](/listings/nagisanzenin--production-grade)

HTML

<a href="/listings/nagisanzenin--production-grade"><img src="/api/listings/nagisanzenin--production-grade/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

# Claude Code — marketplace (recommended)
/plugin marketplace add nagisanzenin/claude-code-plugins
/plugin install production-grade@nagisanzenin

# Or from source
git clone https://github.com/nagisanzenin/production-grade.git
claude --plugin-dir /path/to/production-grade

# Other platforms: see the support matrix above / INSTALL-<PLATFORM>.md

**Requirements:** an agent platform from the matrix above (Claude Code is the reference), Docker & Docker Compose, Git.

Works on existing codebases — brownfield detection auto-maps your project structure.

---

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

plugin.json · Manifest · 5 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. {
  2. "$schema": "https://antigravity.google/schemas/v1/plugin.json",
  3. "name": "production-grade",
  4. "description": "14 AI agents, one install, idea to production: architecture docs, API contracts, tested backend/frontend, security audit, CI/CD pipelines, and documentation — for building, hardening, shipping, and documenting real systems."
  5. }
README.md · README · 505 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. # Production Grade Plugin
  2. <p align="center">
  3. <img src="assets/banner.png" alt="Meet the Production Grade crew — 14 AI agents" width="700">
  4. </p>
  5. <p align="center">
  6. <a href="https://github.com/nagisanzenin/production-grade"><img src="https://img.shields.io/github/stars/nagisanzenin/production-grade?style=social" alt="GitHub stars"></a>
  7. <a href="https://discord.gg/3ux2c5xz"><img src="https://img.shields.io/badge/Discord-Join%20Community-5865F2?logo=discord&logoColor=white" alt="Discord"></a>
  8. <img src="https://img.shields.io/badge/License-MIT-yellow.svg" alt="MIT License">
  9. <img src="https://img.shields.io/badge/version-5.6.1-blue.svg" alt="Version">
  10. <img src="https://img.shields.io/badge/platforms-4-orange.svg" alt="4 platforms">
  11. <img src="https://img.shields.io/badge/agents-14-green.svg" alt="14 agents">
  12. <img src="https://img.shields.io/badge/protocols-9-red.svg" alt="9 protocols">
  13. <img src="https://img.shields.io/badge/execution%20modes-10-purple.svg" alt="10 modes">
  14. </p>
  15. <h3 align="center">14 AI agents, one install, idea to production.</h3>
  16. ```bash
  17. /plugin marketplace add nagisanzenin/claude-code-plugins
  18. /plugin install production-grade@nagisanzenin
  19. ```
  20. > **New in v5.6 — one repo, four platforms.** The same 14 skills now install natively on **Claude Code, OpenAI Codex, OpenCode, and Pi** — one shared core, thin per-platform glue, honest per-platform status. [Support matrix below.](#platforms)
  21. > **New in v5.5 — The Loop Engine.** Agents stop claiming "done" and start proving it. Every stage now loops against a check it can't argue with — typecheck, tests, or the running app clicking its own buttons — until it converges. [How it works.](#the-loop-engine)
  22. <br>
  23. ## Built With This Plugin
  24. **Built something with this plugin? [Open a PR](https://github.com/nagisanzenin/production-grade/pulls) to add your project here.**
  25. | Project | Live | Description |
  26. |---------|------|-------------|
  27. | **PingBase** | [pingbasez.vercel.app](https://pingbasez.vercel.app/) | Free uptime monitoring — get emailed when your website goes down. GitHub OAuth, Stripe billing, Turso DB. |
  28. | **LLM Matrix Arena** | [llm-matrix.vercel.app](https://llm-matrix.vercel.app/) | Browse and compare LLM models across N dimensions. Community-driven voting from real developers — not benchmarks, real opinions. |
  29. | **SkyClaw** | [github.com/nagisanzenin/skyclaw](https://github.com/nagisanzenin/skyclaw) | Cloud-native Rust AI agent runtime. Telegram-native — deploy one binary, paste your API key, and control your server through chat. |
  30. ---
  31. ## Release Timeline
  32. ```
  33. 2026-08-19 v5.6 ●━━━ Omni-platform — native installs on Codex, OpenCode, Pi; platform-adaptation contract; pi payload-harness 14/14
  34. │
  35. 2026-07-02 v5.5.2 ●━━━ Loop-engine robustness — resilient report-artifact writes + orchestrator recovery; deep-smoke requires an adversarial input
  36. │
  37. 2026-07-02 v5.5.1 ●━━━ Loop-engine hardening — trustworthy BUILD-exit deep-smoke gate, worktree-safe scoped oracle-gate, brownfield toolchain adoption
  38. │
  39. 2026-07-02 v5.5 ●━━━ The Loop Engine — oracle-driven iteration: edit-gate hook, TDD pair, functional drive replaces hand-testing
  40. │
  41. 2026-03-07 v5.4 ●━━━ Harmonization — mode-aware autonomy, cross-session enforcement, agent skill loading
  42. │
  43. 2026-03-07 v5.3 ●━━━ Worktree isolation, self-healing gates, cost dashboard
  44. │
  45. 2026-03-07 v5.2 ●━━━ Frontend overhaul — functional-first, design polish, 4 visual style presets
  46. │
  47. 2026-03-07 v5.1 ●━━━ Boundary Safety — 6 patterns for system boundary bugs, from real deployment
  48. │
  49. 2026-03-06 v5.0 ●━━━ Verified & Resilient — receipt enforcement, re-anchoring, adversarial review
  50. │
  51. 2026-03-06 v4.4 ●━━━ Freshness protocol — agents WebSearch to verify volatile data before implementing
  52. │
  53. 2026-03-06 v4.3 ●━━━ Visual identity, pipeline dashboard, gate ceremonies
  54. │
  55. 2026-03-06 v4.2 ●━━━ Adaptive routing, 10 execution modes, everyday SWE work
  56. │
  57. 2026-03-05 v4.1 ●━━━ Engagement modes, scale-driven architecture, adaptive interviews
  58. │
  59. 2026-03-04 v4.0 ●━━━ Two-wave parallelism, internal skill agents, dynamic task generation
  60. │
  61. 2026-03-04 v3.3 ●━━━ Brownfield-safe — works on existing codebases
  62. │
  63. 2026-03-03 v3.2 ●━━━ Auto-update, MECE intent routing, protocol crash fix
  64. │
  65. 2026-03-02 v3.1 ●━━━ Polymath co-pilot — the 14th skill
  66. │
  67. 2026-03-01 v3.0 ●━━━ Full rewrite — Teams/TaskList, 7 parallel points, shared protocols
  68. │
  69. 2026-02-28 v2.0 ●━━━ 13 bundled skills, unified workspace, prescriptive UX
  70. │
  71. 2026-02-24 v1.0 ●━━━ Initial release — autonomous DEFINE>BUILD>HARDEN>SHIP>SUSTAIN
  72. ```
  73. ---
  74. ## The Pipeline
  75. ```
  76. YOU ──→ "Build a SaaS for ..."
  77. │
  78. ▼
  79. ┌─────────────────────────────────────────────────────────────────┐
  80. │ DEFINE │
  81. │ T1 Product Manager ─── BRD, user stories, acceptance criteria│
  82. │ T2 Solution Architect ─ ADRs, API contracts, data models │
  83. │ │
  84. │ ┌─────────────┐ ┌──────────────┐ │
  85. │ │ GATE 1 │ │ GATE 2 │ │
  86. │ │ Requirements│ │ Architecture │ │
  87. │ └─────────────┘ └──────────────┘ │
  88. └─────────────────────────────────────────────────────────────────┘
  89. │
  90. ▼
  91. ┌─────────────────────────────────────────────────────────────────┐
  92. │ BUILD + ANALYZE (Wave A — parallel) │
  93. │ │
  94. │ Backend ──── N agents (1 per service) QA ──── test plan │
  95. │ Frontend ─── N agents (1 per page) Security ── STRIDE │
  96. │ DevOps ──── Dockerfiles + CI skeleton Review ── checklist │
  97. │ SRE ───── SLOs │
  98. └─────────────────────────────────────────────────────────────────┘
  99. │
  100. ▼
  101. ┌─────────────────────────────────────────────────────────────────┐
  102. │ HARDEN (Wave B — parallel against code) │
  103. │ │
  104. │ QA ─────── unit / integration / e2e / performance tests │
  105. │ Security ── code audit + dependency scan (4 parallel phases) │
  106. │ Review ──── arch / quality / performance (adversarial) │
  107. │ DevOps ──── build + push containers │
  108. └─────────────────────────────────────────────────────────────────┘
  109. │
  110. ▼
  111. ┌─────────────────────────────────────────────────────────────────┐
  112. │ SHIP │
  113. │ │
  114. │ DevOps ── IaC + CI/CD ┐ │
  115. │ Remediation ───────────┘ parallel ┌──────────────┐ │
  116. │ SRE ── chaos + capacity ┐ │ GATE 3 │ │
  117. │ Data Scientist ─────────┘ parallel │ Production │ │
  118. │ │ Readiness │ │
  119. │ └──────────────┘ │
  120. └─────────────────────────────────────────────────────────────────┘
  121. │
  122. ▼
  123. ┌─────────────────────────────────────────────────────────────────┐
  124. │ SUSTAIN │
  125. │ │
  126. │ Technical Writer ── API ref + dev guides (parallel) │
  127. │ Skill Maker ─────── 3-5 project-specific reusable skills │
  128. │ Compound Learning ── pipeline insights for next run │
  129. └─────────────────────────────────────────────────────────────────┘
  130. │
  131. ▼
  132. DONE ── receipts verified, agents cleaned up
  133. ```
  134. > **3 gates. 2 waves. 10+ parallel execution points. ~3x faster than sequential.**
  135. > **Every stage loops against an executable oracle until it converges — [The Loop Engine](#the-loop-engine).**
  136. ---
  137. ## The Loop Engine
  138. *New in v5.5 — the reason the agents stop lying about "done."*
  139. Software is never one pass. It's write, run, fail, fix, run again, until a check passes. Through v5.4 the pipeline was mostly feed-forward: each agent did its job, handed the artifact down, and correctness leaned on human review at the gates. v5.5 makes iteration the main path, on a single rule:
  140. > **A task is done when a check it cannot argue with says so — not when the agent claims it.**
  141. That check is an *oracle*: a compiler, a type checker, a test suite, or the running app clicking its own buttons. No oracle, no loop.
  142. ```
  143. ┌──────────────┐
  144. │ PRODUCE │ agent writes / fixes
  145. └──────┬───────┘
  146. ▼
  147. ┌──────────────┐ red ┌────────────────┐
  148. │ ORACLE │ ───────▶ │ DELTA BACK │
  149. │ test·type·app│ │ failing output │
  150. └──────┬───────┘ │ only → re-loop │
  151. green │ └────────┬───────┘
  152. ▼ │
  153. ┌──────────────┐ no progress ×2 │
  154. │ CONVERGED │ ◀─────────────────┘
  155. └──────────────┘ plateau → escalate strategy, not effort
  156. ```
  157. What it adds, concretely:
  158. - **Oracle bootstrap.** Before any parallel wave, the pipeline writes two check scripts for your project: a fast one (typecheck + lint, under 15s) and a full one (tests + build + boot smoke).
  159. - **Enforced by a hook, not a prompt.** A `PostToolUse` hook runs the fast oracle after *every* file edit. Break something and the error lands back in the agent's lap immediately, not at the end.
  160. - **Separated duties.** QA writes failing tests first and owns the `tests/` folder. A coding agent that weakens a test to go green gets flagged with a Critical finding. No grading their own homework.
  161. - **Functional drive.** Before the final gate, an agent boots the real app and drives it — every button, every form, every link. A control that renders but does nothing is a Critical bug. This is the hand-testing replacement.
  162. - **Convergence, not retries.** Each loop tracks a number (failing tests, open findings) and stops when it hits zero or stops improving, shows the trend, then escalates. No retrying the same fix five times.
  163. Fully backward compatible: a pipeline that never needs to loop runs exactly as before. Full design rationale in [docs/LOOPS.md](docs/LOOPS.md).
  164. ---
  165. ## 10 Execution Modes
  166. Not just full builds. The orchestrator reads your request and routes automatically.
  167. ```
  168. ┌──────────────────────────────────────────────────────────────┐
  169. │ │
  170. │ Full Build ████████████████████████ all 14 agents │
  171. │ Feature █████████████ PM+Arch+Eng+QA │
  172. │ Harden ████████ Sec+QA+Review │
  173. │ Ship ██████ DevOps+SRE+DS │
  174. │ Architect ████ Solution Architect │
  175. │ Test ███ QA Engineer │
  176. │ Review ███ Code Reviewer │
  177. │ Document ███ Technical Writer │
  178. │ Optimize █████ SWE+Data Scientist │
  179. │ Explore ███ Polymath │
  180. │ │
  181. └──────────────────────────────────────────────────────────────┘
  182. ```
  183. ```
  184. "Build a SaaS for e-commerce" → Full Build
  185. "Add Stripe billing to my API" → Feature
  186. "Audit this codebase before launch" → Harden
  187. "Set up CI/CD and monitoring" → Ship
  188. "Review this PR for quality" → Review
  189. "Help me think about a fintech app" → Explore
  190. ```
  191. ---
  192. ## The Crew
  193. ```
  194. ┌─────────────────┐
  195. │ ORCHESTRATOR │
  196. │ routes, gates, │
  197. │ receipts │
  198. └────────┬────────┘
  199. │
  200. ┌────────────────────┼────────────────────┐
  201. │ │ │
  202. ┌────▼─────┐ ┌───────▼────────┐ ┌──────▼───────┐
  203. │ DEFINE │ │ BUILD │ │ HARDEN │
  204. │ │ │ │ │ │
  205. │ PM │ │ Software Eng │ │ QA Engineer │
  206. │ Architect│ │ Frontend Eng │ │ Security Eng │
  207. │ │ │ DevOps │ │ Code Review │
  208. └──────────┘ └────────────────┘ └──────────────┘
  209. │
  210. ┌────────────────────┼────────────────────┐
  211. │ │ │
  212. ┌────▼─────┐ ┌───────▼────────┐ ┌──────▼───────┐
  213. │ SHIP │ │ SUSTAIN │ │ ANYTIME │
  214. │ │ │ │ │ │
  215. │ DevOps │ │ Tech Writer │ │ Polymath │
  216. │ SRE │ │ Skill Maker │ │ Data Sci │
  217. └──────────┘ └────────────────┘ └──────────────┘
  218. ```
  219. | # | Agent | Domain | Sole Authority |
  220. |---|-------|--------|:-:|
  221. | 1 | **Orchestrator** | Routes, gates, receipts, re-anchoring | |
  222. | 2 | **Polymath** | Research, ideation, onboarding, translation | |
  223. | 3 | **Product Manager** | BRD, user stories, acceptance criteria | Requirements |
  224. | 4 | **Solution Architect** | ADRs, tech stack, API contracts, data models | Architecture |
  225. | 5 | **Software Engineer** | Handlers, services, repositories, business logic | |
  226. | 6 | **Frontend Engineer** | Design system, components, pages, accessibility | |
  227. | 7 | **QA Engineer** | Unit, integration, e2e, performance tests | |
  228. | 8 | **Security Engineer** | STRIDE, OWASP, PII, dependency scanning | Security |
  229. | 9 | **Code Reviewer** | Architecture conformance, anti-patterns (adversarial) | Code Quality |
  230. | 10 | **DevOps** | Docker, Terraform, CI/CD, containers | Infrastructure |
  231. | 11 | **SRE** | SLOs, chaos engineering, runbooks, capacity | Reliability |
  232. | 12 | **Data Scientist** | LLM optimization, prompt engineering, cost modeling | |
  233. | 13 | **Technical Writer** | API reference, dev guides, architecture docs | |
  234. | 14 | **Skill Maker** | Generates project-specific reusable agent skills | |
  235. ---
  236. ## What Makes It Different
  237. ```
  238. ┌──────────────────────────────────────────────────────────────┐
  239. │ │
  240. │ ORACLE-DRIVEN LOOPS FUNCTIONAL DRIVE │
  241. │ ────────────────── ──────────────── │
  242. │ No oracle, no loop. A loop A separate agent boots the │
  243. │ exits on an executable running app and drives it: │
  244. │ check: tests, typecheck, every button, every form, │
  245. │ the running app. Never on every link. A dead element │
  246. │ "looks done." Converge or is a Critical bug — this is │
  247. │ escalate, never blind-retry. the hand-testing killer. │
  248. │ │
  249. │ RECEIPT ENFORCEMENT RE-ANCHORING │
  250. │ ───────────────── ──────────── │
  251. │ Every agent writes a Orchestrator re-reads specs │
  252. │ JSON receipt as proof. FROM DISK at every phase │
  253. │ No receipt = not done. transition. No context drift │
  254. │ Gate won't open without in multi-hour runs. │
  255. │ verified artifacts. │
  256. │ │
  257. │ ADVERSARIAL REVIEW FRESHNESS PROTOCOL │
  258. │ ────────────────── ────────────────── │
  259. │ Code reviewer assumes Agents detect volatile data │
  260. │ code is WRONG until (model IDs, pricing, CVEs) │
  261. │ proven right. Scales and WebSearch to verify │
  262. │ from critical-only to BEFORE implementing. │
  263. │ hostile break scenarios. │
  264. │ │
  265. │ CONSTRAINT-DRIVEN ARCH ZERO OPEN-ENDED QUESTIONS │
  266. │ ───────────────────── ───────────────────────── │
  267. │ Architecture derived from Every interaction is arrow │
  268. │ YOUR scale, budget, team, keys + Enter. Polymath │
  269. │ compliance — not templates. translates at every gate. │
  270. │ 100 users → monolith. │
  271. │ 10M users → microservices. │
  272. │ │
  273. │ MODE-AWARE AUTONOMY CROSS-SESSION PERSISTENCE │
  274. │ ─────────────────── ───────────────────────── │
  275. │ Express: zero questions, SessionStart hook detects │
  276. │ auto-resolve everything. production-grade projects. │
  277. │ Meticulous: every decision New sessions get a courteous │
  278. │ surfaced. Agent questions prompt: use plugin, work │
  279. │ scale independently of directly, or chat about it. │
  280. │ pipeline gates. Your workflow persists. │
  281. │ │
  282. └──────────────────────────────────────────────────────────────┘
  283. ```
  284. ---
  285. ## Protocol Stack
  286. All 14 agents load the same 9 protocols at startup:
  287. ```
  288. ┌──────────────────────────────────────────────┐
  289. │ Loop Protocol │ ← oracle-driven iteration
  290. ├──────────────────────────────────────────────┤
  291. │ Boundary Safety │ ← system boundary patterns
  292. ├──────────────────────────────────────────────┤
  293. │ Receipt Protocol │ ← proof of completion
  294. ├──────────────────────────────────────────────┤
  295. │ Freshness Protocol │ ← verify volatile data
  296. ├──────────────────────────────────────────────┤
  297. │ Visual Identity │ ← consistent formatting
  298. ├──────────────────────────────────────────────┤
  299. │ Conflict Resolution │ ← sole-authority domains
  300. ├──────────────────────────────────────────────┤
  301. │ Tool Efficiency │ ← dedicated tools > shell
  302. ├──────────────────────────────────────────────┤
  303. │ Input Validation │ ← classify external inputs
  304. ├──────────────────────────────────────────────┤
  305. │ UX Protocol │ ← structured interactions
  306. └──────────────────────────────────────────────┘
  307. ```
  308. ---
  309. ## Engagement Modes
  310. Choose your depth at pipeline start. Propagates to all 14 agents.
  311. ```
  312. Express ░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░ zero agent questions, auto-resolve all
  313. Standard ████░░░░░░░░░░░░░░░░░░░░░░░░░░ 1-2 per skill, subjective/irreversible only
  314. Thorough █████████████░░░░░░░░░░░░░░░░░░ all major decisions surfaced
  315. Meticulous ██████████████████████████████░░ every decision point, full user control
  316. ```
  317. > **3 pipeline gates (BRD, Architecture, Production Readiness) always fire regardless of mode.**
  318. > Agent questions are separate — they scale from zero (Express) to exhaustive (Meticulous).
  319. ---
  320. ## Token-Efficient Architecture
  321. Large skills split into **router + on-demand phases**. Only what's needed loads. Independent phases run as parallel agents with minimal context.
  322. ```
  323. Polymath ─────────── 6 modes onboard | research | ideate | advise | translate | synthesize
  324. Software Engineer ── 5 phases context | implement | cross-cutting | integration | local dev
  325. Frontend Engineer ── 6 phases analysis | functional foundation | components | pages | design polish | testing
  326. Security Engineer ── 6 phases threat model | code audit | auth | data | supply chain | remediation
  327. SRE ─────────────── 5 phases readiness | SLOs | chaos | incidents | capacity
  328. Data Scientist ──── 6 phases audit | LLM optimization | experiments | pipeline | ML infra | cost
  329. Technical Writer ── 4 phases audit | API reference | dev guides | Docusaurus
  330. ```
  331. ---
  332. ## By the Numbers
  333. ```
  334. ┌─────────────────────────────────────────────────────┐
  335. │ │
  336. │ 14 specialized agents │
  337. │ 9 shared protocols │
  338. │ 2 oracle scripts generated per project │
  339. │ 1 fast oracle check after every single edit │
  340. │ 10 execution modes │
  341. │ 10+ parallel execution points │
  342. │ 3 approval gates │
  343. │ 4 engagement modes │
  344. │ ~3x faster than sequential execution │
  345. │ ~45% fewer input tokens from parallelism │
  346. │ 0 open-ended questions — all structured │
  347. │ 11 governing principles │
  348. │ 5 languages: TS, Go, Python, Rust, Java/Kotlin │
  349. │ │
  350. └─────────────────────────────────────────────────────┘
  351. ```
  352. ---
  353. ## Platforms
  354. One repo, one shared core (`skills/` — Agent Skills standard `SKILL.md` — plus the protocols in `skills/_shared/`), thin per-platform glue. Where a platform lacks a tool the skills name, `skills/_shared/platform-adaptation.md` maps the mechanism; the guarantees (structured choices, receipts, gates, re-anchoring, oracle discipline) stay identical.
  355. | Platform | Install | Invoke | Status |
  356. |---|---|---|---|
  357. | **Claude Code** | `/plugin marketplace add nagisanzenin/claude-code-plugins` then `/plugin install production-grade@nagisanzenin` | `/production-grade …` or natural language | Reference platform — hooks, parallel waves, worktrees, everything |
  358. | **OpenAI Codex** | [INSTALL-CODEX.md](INSTALL-CODEX.md) — plugin route or Agent Skills route | `$production-grade …` | Manifests mirror the Claude ones; **not yet verified against a live Codex binary** (honest status in the doc) |
  359. | **OpenCode** | [INSTALL-OPENCODE.md](INSTALL-OPENCODE.md) — clone + one `skills.paths` config entry | natural language | **Verified live on v1.18.4**: all 14 skills discovered (`opencode debug skill`) |
  360. | **Pi** | `pi install git:github.com/nagisanzenin/production-grade` — [INSTALL-PI.md](INSTALL-PI.md) | `/production-grade …` or `/skill:production-grade` | **Verified on pi 0.84.2, 14/14 checks** by the committed [payload-capture harness](experiments/pi-harness/) — skills, template, session-guard extension |
  361. Ambient enforcement varies by host: the SessionStart guard + PostToolUse oracle-gate hooks run on Claude Code (and ride along for Codex); Pi gets the session guard via `extensions/production-grade.ts`; on hosts without edit hooks, agents run the fast oracle themselves after every edit (loop-protocol Rule 7 — the hook is enforcement, not the rule).
  362. ## Installation
  363. ```bash
  364. # Claude Code — marketplace (recommended)
  365. /plugin marketplace add nagisanzenin/claude-code-plugins
  366. /plugin install production-grade@nagisanzenin
  367. # Or from source
  368. git clone https://github.com/nagisanzenin/production-grade.git
  369. claude --plugin-dir /path/to/production-grade
  370. # Other platforms: see the support matrix above / INSTALL-<PLATFORM>.md
  371. ```
  372. **Requirements:** an agent platform from the matrix above (Claude Code is the reference), Docker & Docker Compose, Git.
  373. Works on existing codebases — brownfield detection auto-maps your project structure.
  374. ---
  375. ## FAQ
  376. **Does it write working code?** Yes. Write, build, test, debug, fix. No stubs. No TODOs.
  377. **Existing projects?** Yes. Brownfield detection auto-maps. Run specific modes or full pipeline.
  378. **How do I know it ran everything?** Receipts. JSON proof from every agent, verified at gates.
  379. **Context degrade in long runs?** No. Re-anchoring re-reads from disk at every phase transition.
  380. **Not technical?** Every interaction is multiple choice. Polymath translates at any gate.
  381. ---
  382. ## Contributing
  383. 1. Fork the repo
  384. 2. Create a branch: `git checkout -b feature/your-feature`
  385. 3. Commit changes
  386. 4. Open a Pull Request
  387. **Adding a skill:** Create `skills/your-skill-name/SKILL.md` with `---` frontmatter.
  388. ---
  389. ## Community
  390. Join the Discord to share what you've built, discuss workflows, report bugs, and request features.
  391. <a href="https://discord.gg/3ux2c5xz"><img src="https://img.shields.io/badge/Discord-Join%20Community-5865F2?style=for-the-badge&logo=discord&logoColor=white" alt="Join Discord"></a>
  392. ---
  393. ## Star History
  394. <a href="https://star-history.com/#nagisanzenin/production-grade&Date">
  395. <picture>
  396. <source media="(prefers-color-scheme: dark)" srcset="https://api.star-history.com/svg?repos=nagisanzenin/production-grade&type=Date&theme=dark" />
  397. <source media="(prefers-color-scheme: light)" srcset="https://api.star-history.com/svg?repos=nagisanzenin/production-grade&type=Date" />
  398. <img alt="Star History Chart" src="https://api.star-history.com/svg?repos=nagisanzenin/production-grade&type=Date" />
  399. </picture>
  400. </a>
  401. ---
  402. ## More from the same workshop
  403. Five Claude Code plugins from the same workshop. Most share one habit: *let a deterministic core decide, and never let the producer of work grade it.*
  404. - **[engram](https://github.com/nagisanzenin/engram)** — evidence-based learning engine: first-principles curricula, generation-first tutoring, blind-graded free recall, FSRS-scheduled memory. The receipt-and-assessor discipline in this plugin started there.
  405. - **[effortmining](https://github.com/nagisanzenin/effortmining)** — benchmark-calibrated per-subagent reasoning effort: dispatch the cheapest tier a blind grader still accepts (64.7% fewer output tokens at equal quality, pre-registered). Built *with* this pipeline — its repo carries the receipts.
  406. - **[idiolect](https://github.com/nagisanzenin/idiolect)** — human-voice writing engine: 60+ measured voices plus a deterministic AI-tell scanner and a blind auditor, so text reads like a person, not a model.
  407. - **[less](https://github.com/nagisanzenin/less)** — a minimal comms protocol for Claude: a per-turn hook makes replies answer-first, pick-list-driven, and calm, without touching the work.
  408. ## License
  409. MIT
  410. ---
  411. <p align="center">
  412. <strong>14 agents. 9 protocols. 10 modes. One install.</strong>
  413. </p>

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.