Documentation-first development methodology. The goal is AI-ready documentation - when docs are clear enough, code generation becomes automatic. Triggers on "Build", "Create", "Implement", "Document", or "Spec out". Version 3.5 adds Phase 2.5 Adversarial Review and renames internal verification to Spec Gate (structural completeness). Clarity Gate is now a separate standalone tool for epistemic quality.
How to install stream-coding?
frmoretto/stream-coding
$npx -y skills add frmoretto/stream-coding --skill stream-codingInstalls into the current project.
›Prefer a prompt? Paste this to your agent
Run `npx skills use "https://github.com/frmoretto/stream-coding" --skill "frmoretto/stream-coding"` and follow the generated skill instructions now. Read its complete output, redirecting it to a temporary file first if necessary. Resolve relative paths from the supporting-files directory it provides.
Use the skills in "https://github.com/frmoretto/stream-coding" that are relevant to the current task. Run `npx skills add "https://github.com/frmoretto/stream-coding"` and select the relevant skills, then follow their instructions.
SKILL.md
| 1 | # Stream Coding v3.5 |
| 2 | |
| 3 | **The 10-20x Methodology for AI-Accelerated Software Development** |
| 4 | |
| 5 | [](https://creativecommons.org/licenses/by/4.0/) |
| 6 | |
| 7 | > 📢 **v3.5 Update:** Phase 2.5 Adversarial Review added, internal gate renamed Spec Gate (structural), Clarity Gate is now a standalone epistemic tool. Re-download if using older version. |
| 8 | |
| 9 | > ✅ **This methodology built 7 production modules in 4.5 hours** ([5Levels Case Study](./manifesto/Appendix_C_5Levels_Case_Study.md), Git-verified) |
| 10 | |
| 11 | > *"Stream coding isn't about faster coding. It's about documentation so clear that code writes itself."* |
| 12 | |
| 13 | --- |
| 14 | |
| 15 | ## The Problem: The Velocity Mirage |
| 16 | |
| 17 | AI tools promise 10x productivity. GitHub Copilot, Cursor, Claude Code—they make coding 55% faster. |
| 18 | |
| 19 | **But projects still take the same time to ship.** |
| 20 | |
| 21 | Why? Because faster typing doesn't solve: |
| 22 | - Strategic decisions AI can't make for you |
| 23 | - Context that gets lost between prompts |
| 24 | - Technical debt created at 10x speed |
| 25 | |
| 26 | This gap between task velocity and project velocity is the **Velocity Mirage**. |
| 27 | |
| 28 | --- |
| 29 | |
| 30 | ## The Solution: Stream Coding |
| 31 | |
| 32 | Stream Coding is a documentation-first methodology that makes AI-generated code deterministic. |
| 33 | |
| 34 | **The 40/40/5/10/5 Split:** |
| 35 | - **40%** Strategic Thinking (Phase 1) — Solve hard problems before coding |
| 36 | - **40%** AI-Ready Documentation (Phase 2) — Specs so complete AI has zero decisions |
| 37 | - **5%** Adversarial Review (Phase 2.5) — Different AI attacks the specs before coding |
| 38 | - **10%** Execution (Phase 3) — Code streams out automatically |
| 39 | - **5%** Quality (Phase 4) — Tests and verification |
| 40 | |
| 41 | **Real Results (5Levels Case Study):** |
| 42 | - 7 production modules in 4.5 hours |
| 43 | - 46 intelligence endpoints (77 total backend API) |
| 44 | - Zero bugs in generated code, 21 minutes average per tested module |
| 45 | |
| 46 | *Note: The case study focuses on backend intelligence modules—Stream Coding's sweet spot. For frontend, use the methodology for behavior (components, state, logic) and complement with AI design tools for visuals. See Chapter 4 for details.* |
| 47 | |
| 48 | --- |
| 49 | |
| 50 | ## Quick Start |
| 51 | |
| 52 | ### Option 1: Claude.ai / Claude Desktop |
| 53 | |
| 54 | 1. Download [`stream-coding.skill`](dist/stream-coding.skill) |
| 55 | 2. Go to Settings → Features → Skills → Add |
| 56 | 3. Upload the `.skill` file |
| 57 | 4. Ask Claude: *"Build a user authentication system"* |
| 58 | |
| 59 | ### Option 2: Claude Code |
| 60 | |
| 61 | Clone the repo — Claude Code auto-detects skills in `.claude/skills/`: |
| 62 | |
| 63 | ```bash |
| 64 | git clone https://github.com/frmoretto/stream-coding |
| 65 | cd stream-coding |
| 66 | # Claude Code will automatically detect .claude/skills/stream-coding/SKILL.md |
| 67 | ``` |
| 68 | |
| 69 | Or copy `.claude/skills/stream-coding/` to your project's `.claude/skills/` directory. |
| 70 | |
| 71 | Ask Claude: *"Build a user authentication system"* |
| 72 | |
| 73 | ### Option 3: Claude Projects |
| 74 | |
| 75 | Add [`SKILL.md`](SKILL.md) to project knowledge. Claude will search it when needed, though Skills provide better integration. |
| 76 | |
| 77 | ### Option 4: OpenAI Codex / GitHub Copilot |
| 78 | |
| 79 | Copy the canonical skill to the appropriate directory: |
| 80 | |
| 81 | | Platform | Location | |
| 82 | |----------|----------| |
| 83 | | OpenAI Codex | `.codex/skills/stream-coding/SKILL.md` | |
| 84 | | GitHub Copilot | `.github/skills/stream-coding/SKILL.md` | |
| 85 | |
| 86 | Use [`skills/stream-coding/SKILL.md`](skills/stream-coding/SKILL.md) (agentskills.io format). |
| 87 | |
| 88 | ### Option 5: Manual / Other LLMs |
| 89 | |
| 90 | For Cursor, Windsurf, or other AI tools: |
| 91 | - Extract core principles (Phases, Document Types, Spec Gate) |
| 92 | - Create a condensed version for `.cursorrules` or project settings |
| 93 | - Use the [templates](./templates) and [Spec Gate Checklist](./templates/SPEC_GATE_CHECKLIST.md) as reference |
| 94 | |
| 95 | The methodology is tool-agnostic—only SKILL.md is Claude-optimized. |
| 96 | |
| 97 | --- |
| 98 | |
| 99 | ## Read the Manifesto |
| 100 | |
| 101 | The [`/manifesto`](./manifesto) folder contains the complete methodology: |
| 102 | |
| 103 | | Chapter | Topic | |
| 104 | |---------|-------| |
| 105 | | [Chapter 1](./manifesto/Chapter_01_The_10x_Promise.md) | The Velocity Mirage | |
| 106 | | [Chapter 2](./manifesto/Chapter_02_Why_AI_Cant_Deliver.md) | Why AI Tools Alone Fail | |
| 107 | | [Chapter 3](./manifesto/Chapter_03_The_Methodology_Gap.md) | The Missing Middle | |
| 108 | | [Chapter 4](./manifesto/Chapter_04_What_Is_Stream_Coding.md) | The 5-Phase Methodology | |
| 109 | | [Chapter 5](./manifesto/Chapter_05_The_Perpetual_Stream.md) | Day 2 & The Rule of Divergence | |
| 110 | | [Appendix A](./manifesto/Appendix_A_The_Toolkit.md) | Templates & Checklists | |
| 111 | | [Appendix B](./manifesto/Appendix_B_References.md) | Research & SDD Positioning | |
| 112 | | [Appendix C](./manifesto/Appendix_C_5Levels_Case_Study.md) | 5Levels Case Study (Git-Verified) | |
| 113 | | [Advanced Framework](./manifesto/advanced/Advanced_Framework.md) | Document Architecture (v3.5) | |
| 114 | |
| 115 | ### Use the Templates |
| 116 | |
| 117 | The [`/templates`](./templates) folder contains ready-to-use frameworks: |
| 118 | |
| 119 | | Template | Purpose | |
| 120 | |----------|--------| |
| 121 | | [Strategic Blueprint](./templates/STRATEGIC_BLUEPRINT.md) | Answer the 7 Phase 1 Questions | |
| 122 | | [ADR Template](./templates/ADR_TEMPLATE.md) | Document architecture decisions with rationale | |
| 123 | | [Spec Gate Checklist](./templates/SPEC_GATE_CHECKLIST.md) | The mandatory Phase 2→ |
Related
6 picksType
- azure-messagingTroubleshoot and resolve issues with Azure Messaging SDKs for Event Hubs and Service Bus.Jul 2026
- lark-openapi-explorer飞书/Lark 原生 OpenAPI 探索：从官方文档库中挖掘未经 CLI 封装的原生 OpenAPI 接口。当用户的需求无法被现有 lark-* skill 或 lark-cli 已注册命令满足，需要查找并调用原生飞书 OpenAPI 时使用。Jul 2026
- lark-skill-maker创建 lark-cli 的自定义 Skill。当用户需要把飞书 API 操作封装成可复用的 Skill（包装原子 API 或编排多步流程）时使用。Jul 2026
- implementImplement a piece of work based on a spec or set of tickets.Jul 2026
- supabaseUse when doing ANY task involving Supabase. Triggers: Supabase products (Database, Auth, Edge Functions, Realtime, Storage, Vectors, Cron, Queues); client…Jul 2026
- firebase-basicsProvides foundational setup, authentication, and project management workflows for Firebase using the Firebase CLI.Jul 2026
