Work with Obsidian vaults (plain Markdown notes) and automate via obsidian-cli.
npx skills add bat06140/skills --skill "documentation-creator"
Install specific skill from multi-skill repository
# Description
|
# SKILL.md
name: documentation-creator
description: |
Generate or update comprehensive web documentation for any project. Use for: creating documentation, updating existing docs, generating guides, documenting features with element-specific screenshots.
Triggers: créer une documentation, génère une doc, générer documentation, create docs, generate documentation, document project, write docs, update docs, mettre à jour la doc, documenter le projet.
Supports WordPress plugins, Next.js, PHP, HTML/CSS/JS, Vue.js, monorepos. Uses Playwright Python for element-specific screenshots.
Documentation Creator
Generate or update web documentation adapted to the project's technology stack.
UI Flow Discovery Workflow
IMPORTANT: Use the Playwright session script (
capture_session.py) for ALL screenshots. Do NOT use the agent's built-in browser tool.
Overview
The recommended workflow for capturing documentation screenshots:
- Explore - Browse the UI to discover feature flows
- Analyze - Identify forms, navigation, modals, interactive elements
- Collect Inputs - Ask user for required values (credentials, test data)
- Generate Plan - Create
screenshot_plan.json - Execute - Run
capture_session.py --plan
Phase 1: Project Context (Optional)
Run detect_project_type.py to understand project structure:
python scripts/detect_project_type.py /path/to/project --json
This provides hints about admin URLs and feature locations.
Phase 2: Interactive Flow Discovery
Use capture_session.py interactively to explore the UI:
python scripts/capture_session.py
Discovery commands:
{"cmd": "navigate", "url": "http://localhost/app"}
{"cmd": "get_page_info"}
{"cmd": "get_forms"}
{"cmd": "get_navigation"}
{"cmd": "get_modals"}
Exploration loop:
- Navigate to entry point
- Use
get_navigationto find menu items and action buttons - Click through to each feature page
- Use
get_formsto identify input fields - Use
get_modalsafter clicking buttons that trigger dialogs - Document each flow with its steps and required inputs
Phase 3: Collect Required Values
After identifying flows, ask the user:
I've identified the following UI flows that need documentation:
- Login Flow - Requires: username, password
- Create Widget Flow - Requires: widget_name, widget_type (dropdown)
- Settings Page - No inputs required
Please provide the test values for each flow.
Phase 4: Generate Capture Plan
Create screenshot_plan.json with all commands:
[
{ "cmd": "navigate", "url": "http://localhost/login" },
{ "cmd": "fill", "selector": "#username", "value": "admin" },
{ "cmd": "fill", "selector": "#password", "value": "secret" },
{ "cmd": "click", "selector": "#submit" },
{ "cmd": "wait", "selector": ".dashboard" },
{
"cmd": "screenshot",
"output": "./docs/images/dashboard.png",
"selector": ".main-content"
}
]
Phase 5: Execute Plan
python scripts/capture_session.py --plan /tmp/screenshot_plan.json
Commands Reference
| Command | Parameters | Description |
|---|---|---|
navigate |
url, wait_until? |
Navigate to URL |
find_elements |
selector?, type? |
Find elements (types: input, button, link, form, etc.) |
fill |
selector, value |
Fill an input field |
click |
selector |
Click an element |
select |
selector, value |
Select dropdown option by value |
select_index |
selector, index |
Select dropdown option by index |
wait |
ms?, selector? |
Wait for time or element |
screenshot |
output, selector?, full_page?, padding? |
Take screenshot |
get_page_info |
- | Get page info (url, title, forms count, buttons) |
get_forms |
- | Get all forms with fields, options, submit buttons |
get_navigation |
- | Get menus, tabs, action links for flow discovery |
get_modals |
- | Get visible modals/dialogs with their content |
quit |
- | Close browser and exit |
Quick Start
Create New Documentation
- Detect project type:
python scripts/detect_project_type.py /path/to/project - Explore UI flows: Use interactive session with discovery commands
- Collect user inputs: Ask for credentials and test data
- Generate plan: Create
screenshot_plan.json - Execute:
python scripts/capture_session.py --plan /tmp/screenshot_plan.json
Reference Files
references/project_types.md- Detection patternsreferences/documentation_structure.md- Content templatesreferences/frameworks.md- Framework-specific guidesreferences/ui_flow_analysis.md- UI flow discovery patterns
# Supported AI Coding Agents
This skill is compatible with the SKILL.md standard and works with all major AI coding agents:
Learn more about the SKILL.md standard and how to use these skills with your preferred AI coding agent.