documentation-structure
Documentation architecture for this repository. Use when creating, updating, or reviewing README.md, CONTRIBUTING.md, or docs/ files. Covers separation of concerns, vendor documentation standards, cross-references, and validation.
Works with
---
name: documentation-structure
description: Documentation architecture for this repository. Use when creating, updating, or reviewing README.md, CONTRIBUTING.md, or docs/ files. Covers separation of concerns, vendor documentation standards, cross-references, and validation.
license: MIT
---
# Documentation Structure
This skill defines how documentation is organized and maintained in this repository.
## Core Principles
| Principle | Description |
|-----------|-------------|
| **Separation of Concerns** | README (landing), docs/ (reference), CONTRIBUTING (dev workflow) |
| **Single Source of Truth** | Define once, link everywhere. Never duplicate content. |
| **Hub-and-Spoke** | `docs/README.md` is the central navigation hub |
| **Vendor Isolation** | Each AI platform gets its own directory in docs/ |
## Document Responsibilities
## User vs Developer Content Separation
**CRITICAL RULE:** User documentation must ONLY contain fully automated installation methods. All manual setup belongs in developer documentation.
| Content Type | User Docs | Developer Docs |
|--------------|-----------|----------------|
| Marketplace install | ✓ | ✓ |
| One-command GitHub install | ✓ | ✓ |
| git clone | ✗ | ✓ |
| --plugin-dir | ✗ | ✓ |
| extensions link | ✗ | ✓ |
| JSON config editing | ✗ | ✓ |
| Local path setup | ✗ | ✓ |
| mcp_settings.json | ✗ | ✓ |
### User Documentation (README.md, docs/*/overview.md)
- Installation must be copy-paste simple
- Single command or UI-only steps
- Link to dev docs for manual alternatives
### Developer Documentation (CONTRIBUTING.md, docs/*/*-development.md)
- All manual setup workflows
- Local testing procedures
- Configuration file editing
- Environment setup
### README.md (Landing Page)
**Purpose:** First impression. Get users started quickly.
**Must include:**
- One-line description
- Quick Start (4 steps: Choose → Install → Authenticate → Try)
- Capability tables (what users can do)
- Links to docs/ for details
**Must NOT include:**
- Full API reference (→ docs/mcp/)
- Development workflows (→ CONTRIBUTING.md)
- Detailed architecture (→ docs/)
### docs/ (Reference Documentation)
**Purpose:** Complete reference for users and developers.
**Structure:**
```
docs/
├── README.md # Navigation hub
├── troubleshooting.md # Cross-platform issues
├── getting-started/ # Entry points
│ ├── mcp-setup.md # Generic MCP config
│ └── enterprise.md # Admin requirements
├── claude-code/ # Vendor: Claude Code
├── kiro/ # Vendor: Kiro
├── gemini-cli/ # Vendor: Gemini CLI
└── mcp/ # Protocol reference
├── tools-reference.md
└── tutorials.md
```
### CONTRIBUTING.md (Development Workflow)
**Purpose:** How to modify THIS repository.
**Must include:**
- Clone and local dev setup
- How to test changes locally (`--plugin-dir`, etc.)
- Validation checklists
- PR process
**Must NOT include:**
- Full plugin/power architecture (→ docs/)
- User-facing tutorials (→ docs/)
## Vendor Documentation Standards
Each vendor directory in `docs/` follows this pattern:
### Required Files
| File | Purpose |
|------|---------|
| `overview.md` | What is this integration, why use it, installation |
| `*-development.md` | How to build new plugins/powers/extensions |
| Individual component docs | One file per plugin/power |
### Standard Sections in overview.md
```markdown
## What are [Plugins/Powers/Extensions]?
## Why Use [Plugins/Powers] vs Direct MCP?
## Available [Plugins/Powers]
## Installation
## Authentication
## Related
```
### Vendor-Specific Metadata
| Vendor | Config Format | Location |
|--------|--------------|----------|
| Claude Code | `plugin.json` | `.claude-plugin/plugin.json` |
| Kiro | POWER.md frontmatter | `POWER.md` |
| Gemini CLI | JSON extension | `gemini-extension.json` |
## Cross-Reference Patterns
### Link Format
- **Within docs/**: Use relative paths `[text](../mcp/tools-reference.md)`
- **From README to docs/**: Use `docs/` prefix `[text](docs/claude-code/overview.md)`
- **External links**: Full URLs `[text](https://developers.miro.com)`
### Required "Related" Section
Every doc file should end with a Related section:
```markdown
## Related
- [Overview](overview.md) - Introduction to this integration
- [Tools Reference](../mcp/tools-reference.md) - MCP tool documentation
```
### Reciprocal Links
If doc A links to doc B, doc B should link back to doc A in its Related section.
## Validation Guidelines
Before committing documentation changes:
### Content Checks
- [ ] No duplicated content (link instead)
- [ ] Correct document owns the content (README vs docs/ vs CONTRIBUTING)
- [ ] All sections present per vendor standards
### Link Checks
- [ ] All internal links resolve
- [ ] Related sections have reciprocal links
- [ ] External links use HTTPS
### Format Checks
- [ ] Code blocks have language specified
- [ ] Tables have consistent formatting
- [ ] Collapsibles have matching tags
## See Also
- `references/patterns.md` - Formatting patterns (tables, collapsibles, code blocks)More Writing & Documentation skills
paper-context-resolver
lllllllama/rigorpilot-skills
Rigor Paper Context helper for README-first deep learning repo reproduction. Use only when the README and repository files leave a narrow reproduction-critical gap and the task is to resolve a specific paper detail such as dataset split, preprocessing, evaluation protocol, checkpoint mapping, or runtime assumption from primary paper sources while recording conflicts. Do not use for general paper summary, repo scanning, environment setup, command execution, title-only paper lookup, or replacing README guidance by default.
repo-intake-and-plan
lllllllama/rigorpilot-skills
Rigor Intake helper for README-first deep learning repo reproduction. Use when the task is specifically to scan a repository, read the README and common project files, extract documented commands, classify inference, evaluation, and training candidates, and return the smallest trustworthy reproduction plan to the main orchestrator. Do not use for environment setup, asset download, command execution, final reporting, paper lookup, or end-to-end orchestration.
minimal-run-and-audit
lllllllama/rigorpilot-skills
Rigor Run skill for README-first deep learning repo reproduction. Use when the task is specifically to capture or normalize evidence from the selected smoke test or documented inference or evaluation command and write standardized `repro_outputs/` files, including patch notes when repository files changed. Do not use for training execution, initial repo intake, generic environment setup, paper lookup, target selection, hidden scientific-meaning changes, or end-to-end orchestration by itself.

