f2s-doc-arch

Generate project architecture documentation drafts from user descriptions or codebase scans.

48|6|Updated Feb 11, 2026
One-click install
npx skills add https://github.com/double-coding-lab/Flow2Spec --skill f2s-doc-arch-double-coding-lab
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: f2s-doc-arch
Source: https://github.com/double-coding-lab/Flow2Spec/tree/main/.cursor/skills/f2s-doc-arch
Command: npx skills add https://github.com/double-coding-lab/Flow2Spec --skill f2s-doc-arch-double-coding-lab

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? When starting documentation for a project, teams often lack a written architecture overview, forcing AI coding agents and new contributors to rediscover module boundaries, entry points, and tech stacks on every session. This Skill produces a first-draft architecture document from a text description, an existing document, or a codebase scan. ## Core Features & Use Cases - Draft generation from multiple inputs: Accepts a plain-text description, a local document path (e.g., README.md), or falls back to scanning the project structure with explicit user confirmation. - Orchestrated sub-agent scanning: Supports single-round parallel (B mode) or multi-round corrective (C mode) sub-agent scanning with a strict YAML delivery schema for large or monorepo codebases. - Split recommendations for large features: Appends a split-suggestion section when the codebase exceeds roughly 5000 lines, 20 files, or 3 independent responsibility domains. - Use Case: A developer initializes Flow2Spec on a new repository and runs this Skill with the README as input to produce .Knowledge/stock-docs/架构说明_初稿.md, which later flows into f2s-doc-final and f2s-kb-build for knowledge-base routing. ## Quick Start Ask the agent to run f2s-doc-arch with your README path or a short text description of the project to generate an architecture draft in .Knowledge/stock-docs/.

Frequently Asked Questions about f2s-doc-arch

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I generate a project architecture document with AI?▼

Run f2s-doc-arch with either a plain-text description or a path to an existing document like README.md. The Skill writes a draft architecture overview to .Knowledge/stock-docs/架构说明_初稿.md by default, or to a custom output path you provide.

What happens if I run architecture generation without any input?▼

The Skill falls back to scanning the project directory structure, package.json, and entry files to infer modules and architecture. It must first ask for explicit confirmation because scan-based drafts are not quality-guaranteed and may miss key points.

What is the difference between f2s-doc-arch and f2s-kb-add?▼

f2s-doc-arch only produces architecture description drafts and never writes final documents or builds the knowledge base. f2s-kb-add handles the full pipeline of parsing completed capabilities from multiple files into the knowledge base in one pass.

Can the draft go directly into the knowledge base build step?▼

No. The workflow enforces that drafts must first pass through f2s-doc-final to produce a standardized final document, and only then can f2s-kb-build consume the final version. The only exception is when the user explicitly requests skipping and the draft already meets the final template.

When should a large feature be split into multiple architecture documents?▼

Split when the source exceeds roughly 5000 lines or 20 files, when three or more independent responsibility domains are identifiable, or when the user mentions multiple submodules. The Skill appends a split-suggestion table recommending separate focused drafts per topic.