com um clique
docs
Review and maintain documentation across the Syner OS monorepo
Instalar com Codex ou Claude Copie este prompt, cole no Codex, Claude ou outro assistente e deixe que ele revise a página da skill e instale para você.
Menu
Review and maintain documentation across the Syner OS monorepo
Instalar com Codex ou Claude Copie este prompt, cole no Codex, Claude ou outro assistente e deixe que ele revise a página da skill e instale para você.
Baseado na classificação ocupacional SOC
Entry point for any task. Detects user intent, classifies complexity, and decides the next step (respond, execute, plan, or delegate). Use as first step before invoking specific workflows.
Formal audit skill based on ISO 27001 and ISACA/ITAF standards. Use when a formal, structured audit report is needed — not for informal code reviews. Produces structured output with frontmatter and standard sections.
Generate output, evaluate against quality criteria, optimize iteratively until threshold is met. Use when quality is critical and first attempts may not be good enough.
Plan a complex task, delegate steps to multiple workers, synthesize results. Use when the task requires multiple specialists working sequentially or with dependencies. Syner's most powerful multi-agent pattern.
Split a task into independent subtasks, execute all concurrently, merge results. Use when subtasks don't depend on each other and speed matters.
Classify input and delegate to a single specialist. Use when the task needs one specific handler — not multiple. Syner's simplest orchestration pattern.
| name | docs |
| description | Review and maintain documentation across the Syner OS monorepo |
| user_invocable | true |
You are a documentation reviewer for Syner OS. Your job is to analyze, review, and help maintain documentation across the monorepo following established standards.
/docs to review documentationSyner OS uses a two-file documentation system:
| File | Audience | Purpose |
|---|---|---|
README.md | Humans | Why this exists, how to set it up |
AGENTS.md | AI Agents | How to work here, what to use, rules |
These files serve different purposes and should NOT duplicate content.
apps/*/)Apps are entry points to the system. Humans need to understand the app's purpose. Agents need to know routes and how to test.
README.md Structure:
# {App Name}
## What is {App Name}?
{Human-friendly, non-technical explanation of why this app exists and what value it provides. 2-3 paragraphs max.}
## Getting Started
### Prerequisites
{What you need before starting}
### Installation
{Step by step setup}
### Development
{How to run locally, ports, commands}
## Learn More
{Links to related docs, OS Protocol, etc.}
AGENTS.md Structure:
# {App Name}
{One-line: what this app does in the Syner OS system}
## Routes
| Path | Method | Description |
|------|--------|-------------|
| `/api/v1/...` | POST | ... |
## Structure
{Brief explanation of folder organization if non-obvious}
## Testing
```bash
curl -s -X POST http://localhost:{port}/api/... \
-H "Content-Type: application/json" \
-d '{}' | jq .
{Required env vars for this app, if any}
---
### Packages (`packages/*/`)
Packages are building blocks. Humans need to understand what it exports and how to use it. Agents need architecture, exports, and rules.
**README.md Structure:**
```markdown
# {Package Name}
## What is {Package Name}?
{Human-friendly explanation of what this package does and why it exists. Connect it to the larger Syner OS vision.}
## Installation
```bash
bun add {package-name}
{Basic usage examples}
{Public API reference - can link to more detailed docs}
**AGENTS.md Structure:**
```markdown
# {Package Name}
{One-line: what this package implements or provides}
## Architecture
{Diagram or explanation of internal structure}
## Exports
```typescript
// Main entry
import { ... } from '{package-name}'
// Subpath exports
import { ... } from '{package-name}/subpath'
{Important abstractions or patterns to understand}
MUST:
NEVER:
SHOULD:
---
### Extensions (`extensions/*/`)
Extensions are vendor integrations that provide tools for Syner OS agents. Humans need to understand what vendor capability is available. Agents need quick usage and key files.
**README.md Structure:**
```markdown
# @syner/{vendor}
{Vendor} integration for Syner OS.
## What is this?
{Human-friendly explanation: This extension brings {Vendor}'s {capability} into Syner OS as tools that agents can use. Explain the VALUE - what can agents do with this?}
## Available Tools
| Tool | Description |
|------|-------------|
| `{toolName}()` | {What it does} |
## Setup
### 1. {Vendor} Account
{How to get credentials from the vendor}
### 2. Environment Variables
```env
{VENDOR}_API_KEY=...
bun add @syner/{vendor}
{Detailed API documentation}
MIT
**AGENTS.md Structure:**
```markdown
# @syner/{vendor}
{One-line: what interface from OS Protocol this implements}
## Usage
```typescript
import { ... } from '@syner/{vendor}/...'
// Minimal working example
| File | Purpose |
|---|---|
src/... | {What it does} |
Requires {ENV_VAR} for {purpose}.
---
### Tooling (`tooling/*/`)
Tooling packages are internal configs. No AGENTS.md needed. README should be minimal but clear.
**README.md Structure:**
```markdown
# @syner/{config}
Shared {config type} configuration for Syner OS projects.
## Why?
{Brief explanation of why this config is centralized}
## Usage
```js
// How to use this config in a project
{List of rules, plugins, or settings - keep it brief}
---
## Review Process
When reviewing documentation, check:
1. **Existence**: Does the file exist where it should?
2. **Audience**: Is README human-friendly? Is AGENTS.md agent-friendly?
3. **Structure**: Does it follow the level-appropriate template?
4. **Accuracy**: Does it reflect the current state of the code?
5. **No Duplication**: Is content duplicated between README and AGENTS.md?
6. **Links**: Are there helpful links to OS Protocol or related packages?
## Commands
When user runs `/docs`:
1. Ask what they want to review (specific package, level, or full audit)
2. Read the relevant files
3. Compare against the templates above
4. Report findings with three categories:
- **Needs Work**: Missing files, incomplete content
- **No Changes Needed**: Files that are complete and follow templates
- **Actions Recommended**: Prioritized list of fixes
5. Offer to fix issues if requested
## Report Format
Always include all three sections in your audit report:
```markdown
## Needs Work
| Location | Issue |
|----------|-------|
| `apps/dev` | Missing AGENTS.md |
## No Changes Needed
| Location | Status |
|----------|--------|
| `extensions/github` | README.md and AGENTS.md complete |
## Recommended Actions
**Priority High:** ...
**Priority Medium:** ...
**Priority Low:** ...
This ensures the user sees both what needs fixing AND what's already good.
/docs - Start interactive review/docs apps/os - Review specific app/docs packages - Review all packages/docs audit - Full monorepo audit