A curated collection of Claude Code plugins providing slash commands for design-to-code workflows, context management, multi-agent orchestration, development automation, and visual browser inspection.
- Run
/plugin - Select Add Marketplace
- Enter marketplace source:
arevlo/claude-code-workflows - Select which plugins to install
# Add marketplace
/plugin marketplace add arevlo/claude-code-workflows
# Install specific plugins
/plugin install arevlo:context@claude-code-workflows
/plugin install arevlo:design@claude-code-workflows
/plugin install arevlo:dev@claude-code-workflows
/plugin install arevlo:obsidian@claude-code-workflows
/plugin install arevlo:swarm@claude-code-workflows# Update marketplace
/plugin marketplace update claude-code-workflows
# Then browse and install any new plugins
/plugin
# → Browse and Install PluginsSession context and plan management with multi-source storage.
Save and load Claude Code session context to multiple destinations including local storage, Notion, GitHub Issues, and project docs.
| Command | Description |
|---|---|
/arevlo-context:save-context |
Save current session to local, Notion, GitHub, or docs |
/arevlo-context:load-context <query> |
Search and load prior context from multiple sources |
/arevlo-context:context-status |
Check current context state and recent saves |
/arevlo-context:checkpoint [label] |
Quick checkpoint for mid-task saves (faster than save-context) |
/arevlo-context:plan |
Load, save, or browse Claude Code plans |
| Source | Description | Location |
|---|---|---|
| Session Transcripts | Prior Claude Code sessions (load only) | ~/.claude/projects/ |
| Swarm checkpoints | Auto checkpoints from /auto | .claude/swarm/progress/ |
| Claude Plans | Saved plans | ~/.claude/plans/ |
| Notion | Persistent storage | _clawd database |
| GitHub Issue | Issue tracking in current repo | Creates GitHub issue |
| Docs folder | Project documentation | ./docs/context/ |
| Local /tmp | Quick, ephemeral saves | /tmp/claude-contexts/ |
Session Transcripts are automatically saved by Claude Code for every session. Use /load-context to browse prior sessions with AI-generated summaries, branch info, and message counts.
# End of session
/arevlo-context:save-context
# → Select destination (local, Notion, GitHub, docs)
# → Choose tag type (Context, Summary, Spec, Reference)
# → Auto-generates comprehensive summary
# Start of session
/arevlo-context:load-context "authentication implementation"
# → Choose source to search
# → Select and load previous context
# Mid-task checkpoint
/arevlo-context:checkpoint "before-refactor"
# → Fast save without prompts
# Check status
/arevlo-context:context-status
# → Shows recent saves and checkpoints
# Continue a plan
/arevlo-context:plan
# → Browse and load recent plans- Optional: Notion MCP server for Notion destinations
- Optional: GitHub CLI (
gh) for GitHub Issue destinations
Figma Make design-to-code workflow integration.
Create and manage prompts/changelogs in Notion for Figma Make, enabling design-to-code workflows with proper documentation and iteration tracking.
| Command | Description |
|---|---|
/arevlo-design:make |
Create or update prompts/changelogs in Notion for Figma Make |
1. Create Spec
↓
/arevlo-context:save-context (tag: Spec)
↓
2. Create Prompt
↓
/arevlo-design:make (references saved spec)
↓
3. Figma Make Implementation
↓
4. Iteration (repeat /make for updates)
When creating prompts for Figma Make:
- Specify full file paths (e.g.,
src/components/notcomponents/) - Note file sizes if large (>30KB)
- List ALL files that need modification together
- Provide exact code snippets to insert
- Include verification steps
- Link to related context/specs in Notion
- Notion MCP server configured in Claude Code
- Notion database for storing prompts (default name:
_make)
Git workflows, PR management, and development automation.
Streamline git operations with conventional commits, PR descriptions, code reviews, and release management.
| Command | Description |
|---|---|
/arevlo-dev:commit |
Full commit workflow with branch management and PR updates |
/arevlo-dev:pr-describe |
Generate/update PR description from all changes since main |
/arevlo-dev:pr-review |
Request AI code review via GitHub comment (@claude or @codex) |
/arevlo-dev:release <version> |
Create semver release with tag and GitHub release notes |
/arevlo-dev:kill-port <port> |
Check and kill processes using specific ports |
/arevlo-dev:resolve |
Interactive merge conflict resolution with AI assistance |
# After making code changes
/arevlo-dev:commit
# What it does:
# 1. Checks if branch PR was already merged (prevents lost commits)
# 2. Detects multi-repo scenarios
# 3. Shows what will be committed
# 4. Generates conventional commit message
# 5. Updates PR description if PR exists
# 6. Pushes with confirmationtype: brief description (50 chars max)
Types:
feat:- New featurefix:- Bug fixdocs:- Documentationstyle:- Formattingrefactor:- Code restructuretest:- Testschore:- Maintenance
# Patch version (1.6.0 → 1.6.1)
/arevlo-dev:release patch
# Minor version (1.6.0 → 1.7.0)
/arevlo-dev:release minor
# Major version (1.6.0 → 2.0.0)
/arevlo-dev:release major
# Explicit version
/arevlo-dev:release 2.1.0/arevlo-dev:resolve
# Interactive workflow:
# 1. Detects conflicts in current repo
# 2. Shows each conflict with context
# 3. Offers resolution strategies
# 4. Implements chosen resolution
# 5. Continues to next conflict- GitHub CLI (
gh) installed and authenticated - Git repository with remote configured
Capture screenshots and context to Obsidian Zettelkasten vault as fragment notes.
Quickly save screenshots with analysis and context to your Obsidian vault, supporting the Zettelkasten method of knowledge management. Also save external links to note tables.
| Command | Description |
|---|---|
/obsidian:capture [folder] |
Capture screenshot with context and create fragment note |
/obsidian:save-link <url> [note-path] |
Save external link to note's external links table |
Capture Screenshot:
# Share a screenshot in chat (Slack, Figma, browser, etc.)
[Paste screenshot]
https://life-in-flow.slack.com/archives/...
# Capture it
/obsidian:capture flow
# What it does:
# 1. Analyzes the screenshot content
# 2. Asks for topic/title (e.g., "property-filters-ui")
# 3. Copies screenshot to {vault}/{folder}/_attachments/{topic}.png
# 4. Creates fragment note with embedded image and analysis
# 5. Includes external links (Slack, Figma, etc.)
# 6. Returns note pathSave External Links:
# Save a link with prompts
/obsidian:save-link https://anthropic.com/research/constitutional-ai
# Or specify the note path directly
/obsidian:save-link https://example.com/article ai/outlinks
# What it does:
# 1. Prompts for link title/description
# 2. Prompts for note path (if not provided)
# 3. Optionally asks for tags and notes
# 4. Saves to external links table in the note
# 5. Creates note if it doesn't exist
# 6. Auto-adds source tag from domain (e.g., "source/anthropic")Fragments are temporary captures in the Zettelkasten method:
- Capture - Quickly save ideas/screenshots (this plugin)
- Process - Review and convert to atomic primitives
- Connect - Link primitives to build knowledge graph
Use mcp__obsidian-zettelkasten__process_fragment to convert fragments to primitives later.
- Design discussions - Capture Figma screenshots with Slack context
- Code reviews - Save code snippets with PR comments
- Meeting notes - Capture whiteboard photos with decisions
- Bug reports - Screenshot error states with context
- Obsidian MCP server configured and running
- Obsidian vault accessible at configured path
- Screenshots pasted in Claude Code (auto-cached)
Multi-agent orchestration with autonomous workflows and ACE context management.
Enable parallel code analysis with specialized AI agents and autonomous research-plan-implement workflows for complex tasks.
| Command | Description |
|---|---|
/arevlo-swarm:auto <goal> |
Autonomous workflow - research, plan, implement with checkpoints |
/arevlo-swarm:auto --resume |
Resume from most recent checkpoint |
/arevlo-swarm:swarm [preset] |
Start multi-agent swarm with preset or auto-detect |
/arevlo-swarm:spawn <agent> |
Spawn a single background agent for specific analysis |
/arevlo-swarm:hive |
Check status and findings from all running agents |
/arevlo-swarm:health |
Check context health metrics and alerts |
/arevlo-swarm:sync |
Consolidate findings into prioritized action items |
/arevlo-swarm:fix |
Interactive mode to address swarm findings one by one |
/arevlo-swarm:stop [agent] |
Stop one or all running agents |
/arevlo-swarm:checkpoint [label] |
Create checkpoint during /auto sessions |
/arevlo-swarm:handoff |
Prepare handoff when approaching context limits |
/arevlo-swarm:resume |
Resume from a prior checkpoint or auto phase |
/arevlo-swarm:compact |
Export session then compact working memory |
For complex tasks requiring research, planning, and implementation:
/arevlo-swarm:auto "add user authentication with OAuth"
# Phase 1: Research
# - Explores codebase in isolated context
# - Produces structured findings
# - Saves checkpoint
# Phase 2: Plan
# - Creates phase-by-phase implementation blueprint
# - Waits for human approval (high leverage point)
# - Saves plan checkpoint
# Phase 3: Implement
# - Executes plan phase by phase
# - Auto-saves checkpoints after each phase
# - Monitors context usage
# Resume interrupted session
/arevlo-swarm:auto --resumeFor code review and quality analysis:
# Auto-detect project and recommend agents
/arevlo-swarm:swarm
# Or use a preset
/arevlo-swarm:swarm review
# Check agent status and findings
/arevlo-swarm:hive
# Consolidate all findings
/arevlo-swarm:sync
# Fix issues interactively
/arevlo-swarm:fix
# Stop when done
/arevlo-swarm:stop| Preset | Agents | Use Case |
|---|---|---|
review |
reviewer, simplifier, comment-analyzer | General code review |
quality |
reviewer, type-analyzer, test-analyzer | Code quality & correctness |
security |
silent-hunter, reviewer | Error handling & safety |
cleanup |
simplifier, comment-analyzer | Tech debt reduction |
full |
All agents | Comprehensive analysis |
figma |
reviewer, type-analyzer, silent-hunter | Figma plugin development |
Analysis Agents:
| Agent | Focus | Best For |
|---|---|---|
reviewer |
Code quality, patterns, best practices | Any project |
simplifier |
Complexity reduction, DRY violations | Refactoring, brownfield |
type-analyzer |
TypeScript type safety | TypeScript projects |
silent-hunter |
Unhandled async, silent failures | Async code, error handling |
comment-analyzer |
TODOs, FIXMEs, documentation debt | Cleanup, documentation |
test-analyzer |
Test coverage and quality | Testing improvements |
Orchestration Agents:
| Agent | Focus | Best For |
|---|---|---|
researcher |
Deep codebase exploration | /auto research phase |
coordinator |
Prioritize and group issues | Multi-agent coordination |
/arevlo-swarm:health
# Shows metrics:
# - File reads (threshold: 50)
# - Code edits (threshold: 30)
# - Message weight (threshold: 30)
# - Alert level (GOOD → WATCH → WARNING → CRITICAL)Automatic Features:
- Health metrics tracked via hooks
- Checkpoints saved at 40%, 60%, 70% thresholds
- Emergency saves before auto-compact (80%)
- Graceful completion when nearing limits
Agents communicate through .claude/swarm/:
.claude/swarm/
├── research/ # /auto Phase 1 outputs
├── plans/ # /auto Phase 2 outputs
├── progress/ # Checkpoints (auto + agents)
├── reports/ # Agent findings (timestamped)
├── issues/ # Extracted issues
├── context/ # Shared state
├── logs/ # Agent logs
├── pids/ # Process IDs
└── sync/ # Consolidated reports
/arevlo-swarm:fix
# Workflow:
# 1. Shows issue and current code
# 2. Offers: Fix, Skip, or View more context
# 3. Implements fix and shows diff
# 4. Continues to next issue
# 5. Tracks progress through queue- Claude Code with plugin support
--dangerously-skip-permissionsenabled for background agents- Platforms: macOS, Linux, Windows (PowerShell, Git Bash, or WSL)
# 1. Create and save spec
# (Work with Claude Code to create detailed spec)
/arevlo-context:save-context
# → Select: Notion
# → Tag: Spec
# → Title: "User authentication flow"
# 2. Create Figma Make prompt
/arevlo-design:make
# → References saved spec
# → Creates implementation prompt
# 3. (Figma Make implements the design)
# 4. Review with swarm
/arevlo-swarm:swarm review
# 5. Check findings
/arevlo-swarm:hive
# 6. Fix issues
/arevlo-swarm:fix
# 7. Commit changes
/arevlo-dev:commit
# 8. Create release
/arevlo-dev:release minor# Start autonomous workflow
/arevlo-swarm:auto "implement OAuth authentication with Google and GitHub"
# → Research phase (explores codebase)
# → Plan phase (creates blueprint, waits for approval)
# → Implement phase (executes with checkpoints)
# If interrupted, resume
/arevlo-swarm:auto --resume
# Check context health anytime
/arevlo-swarm:health
# Final commit
/arevlo-dev:commit# 1. Start comprehensive analysis
/arevlo-swarm:swarm full
# 2. Monitor progress
/arevlo-swarm:hive
# 3. Consolidate findings
/arevlo-swarm:sync
# 4. Address issues interactively
/arevlo-swarm:fix
# 5. Save findings for later
/arevlo-context:save-context
# → Tag: Summary
# → Title: "Q4 2024 code audit"
# 6. Stop agents
/arevlo-swarm:stop# During work - checkpoint frequently
/arevlo-context:checkpoint "before-database-refactor"
# Mid-session - check context health
/arevlo-swarm:health
# When warned - save and compact
/arevlo-swarm:compact
# End of session - comprehensive save
/arevlo-context:save-context
# Next session - resume
/arevlo-context:load-context "database refactor"- Claude Code with plugin support
Required for: arevlo-context (Notion destinations), arevlo-design
Add to ~/.claude/.claude.json under mcpServers:
{
"mcpServers": {
"notion": {
"type": "http",
"url": "https://mcp.notion.com/mcp"
}
}
}Authenticate via Notion when prompted.
Notion Database Setup:
For arevlo-context, create database named _clawd with:
- Name (title) - Context title
- Project (select) - Project name
- Tags (multi-select) - Context, Summary, Spec, Reference
For arevlo-design, create database named _make with:
- Name (title) - Prompt title
- Type (select) - Prompt, Changelog
- Project (select) - Project name
Required for: arevlo-dev
# macOS
brew install gh
# Ubuntu/Debian
sudo apt install gh
# Windows
winget install GitHub.cli
# Authenticate
gh auth login- macOS/Linux: bash or zsh (default)
- Windows: PowerShell (default on Windows 10+), Git Bash, or WSL
- Note: cmd.exe is not supported
- Permissions:
--dangerously-skip-permissionsenabled
Update database names and data source IDs in command files:
arevlo:context:
- Default database:
_clawd - Update in:
commands/save-context.md
arevlo:design:
- Default database:
_make - Update in:
commands/make.md
Customize agent focus by editing files in:
plugins/arevlo-swarm/agents/reviewer.mdplugins/arevlo-swarm/agents/silent-hunter.mdplugins/arevlo-swarm/agents/type-analyzer.md- etc.
Add custom presets to plugins/arevlo-swarm/commands/swarm.md:
{
"agents": ["reviewer", "custom-agent"],
"watch": "src/**/*.{ts,tsx}",
"focus": "Your specific focus area"
}See docs/ADDING_PLUGINS.md for how to contribute new plugins to this marketplace.
Plugins follow semantic versioning. Check the marketplace for updates:
/plugin marketplace update claude-code-workflowsCurrent versions:
- arevlo:context: 1.8.1
- arevlo:design: 1.8.1
- arevlo:dev: 1.8.1
- arevlo:obsidian: 1.2.0
- arevlo:swarm: 1.8.1
Contributions welcome! Please:
- Follow existing plugin structure
- Include comprehensive README in plugin directory
- Add examples and workflow documentation
- Test thoroughly before submitting PR
MIT