You're Invited:Meet the Socket Team at RSAC and BSidesSF 2026, March 23–26.RSVP
Socket
Book a DemoSign in
Socket

pro-workflow

Package Overview
Dependencies
Maintainers
1
Versions
5
Alerts
File Explorer

Advanced tools

Socket logo

Install Socket

Detect and block malicious and high-risk dependencies

Install

pro-workflow

Battle-tested Claude Code workflows with agent teams, smart commit, insights, and searchable learnings

Source
npmnpm
Version
1.2.0
Version published
Weekly downloads
35
-77.85%
Maintainers
1
Weekly downloads
 
Created
Source

Pro Workflow

GitHub stars npm version

Battle-tested Claude Code workflows from power users. Self-correcting memory, parallel worktrees, wrap-up rituals, and the 80/20 AI coding ratio.

v1.2.0: Scout agent, /replay, /handoff, drift detection, adaptive quality gates, and correction heatmap!

If this helps your workflow, please give it a star!

What's New in v1.2.0

  • Scout Agent: Confidence-gated exploration — scores readiness (0-100) before implementation, auto-gathers missing context
  • /replay: Surface relevant past learnings before starting a task — your SQLite-powered coding muscle memory
  • /handoff: Generate structured session handoff documents for seamless continuation in the next session
  • Drift Detection: Hook that tracks your original intent and warns when you've strayed from the goal
  • Adaptive Quality Gates: Gates adjust based on your correction history — high correction rate = tighter gates, low rate = relaxed gates
  • Correction Heatmap: /insights heatmap shows which categories and projects get corrected most, with hot/cold learning analysis

Previous: v1.1.0

  • Smart /commit with quality gates, /insights analytics, agent teams, persistent SQLite storage with FTS5

The Core Idea

"80% of my code is written by AI, 20% is spent reviewing and correcting it." — Karpathy

This skill optimizes for that ratio. Every pattern reduces correction cycles.

Patterns

PatternWhat It Does
Self-Correction LoopClaude learns from your corrections automatically
Parallel WorktreesZero dead time - work while Claude thinks
Wrap-Up RitualEnd sessions with intention, capture learnings
Split MemoryModular CLAUDE.md for complex projects
80/20 ReviewBatch reviews at checkpoints
Model SelectionOpus+Thinking for one-shot accuracy
Context DisciplineManage your 200k token budget
Learning LogAuto-document insights

Installation

# Add marketplace
/plugin marketplace add rohitg00/pro-workflow

# Install plugin
/plugin install pro-workflow@pro-workflow

Or via CLI:

claude plugin marketplace add rohitg00/pro-workflow
claude plugin install pro-workflow@pro-workflow

Build with SQLite Support

After installation, build the TypeScript for persistent storage:

cd ~/.claude/plugins/*/pro-workflow  # Navigate to plugin directory
npm install && npm run build

This creates the SQLite database at ~/.pro-workflow/data.db.

Or load directly

claude --plugin-dir /path/to/pro-workflow

Manual Setup

git clone https://github.com/rohitg00/pro-workflow.git /tmp/pw
cp -r /tmp/pw/templates/split-claude-md/* ./.claude/
cp -r /tmp/pw/commands/* ~/.claude/commands/
cp -r /tmp/pw/hooks/* ~/.claude/

Minimal (Just add to CLAUDE.md)

## Pro Workflow

### Self-Correction
When corrected, propose rule → add to LEARNED after approval.

### Planning
Multi-file: plan first, wait for "proceed".

### Quality
After edits: lint, typecheck, test.

### LEARNED

Commands

After plugin install, commands are namespaced:

CommandPurpose
/pro-workflow:wrap-upEnd-of-session checklist
/pro-workflow:learn-ruleExtract correction to memory (file-based)
/pro-workflow:parallelWorktree setup guide
/pro-workflow:learnClaude Code best practices & save learnings
/pro-workflow:searchSearch learnings by keyword
/pro-workflow:listList all stored learnings
/pro-workflow:commitSmart commit with quality gates and code review
/pro-workflow:insightsSession analytics, learning patterns, and correction heatmap
/pro-workflow:replaySurface past learnings for current task
/pro-workflow:handoffGenerate session handoff document for next session

Database Features

Persistent Learnings

Learnings are stored in SQLite with FTS5 full-text search:

~/.pro-workflow/
└── data.db    # SQLite database with learnings and sessions

Search Examples

/search testing           # Find all testing-related learnings
/search "file paths"      # Exact phrase search
/search git commit        # Multiple terms

Learning Categories

  • Navigation (file paths, finding code)
  • Editing (code changes, patterns)
  • Testing (test approaches)
  • Git (commits, branches)
  • Quality (lint, types, style)
  • Context (when to clarify)
  • Architecture (design decisions)
  • Performance (optimization)
  • Claude-Code (sessions, modes, CLAUDE.md, skills, subagents, hooks, MCP)
  • Prompting (scope, constraints, acceptance criteria)

Hooks

Automated enforcement of workflow patterns.

HookWhenWhat
PreToolUseBefore editsTrack edit count, quality gate reminders
PreToolUseBefore git commit/pushRemind about quality gates, wrap-up
PostToolUseAfter code editsCheck for console.log, TODOs, secrets
PostToolUseAfter testsSuggest [LEARN] from failures
UserPromptSubmitEach promptDrift detection — warns when straying from original intent
SessionStartNew sessionLoad learnings from database
StopEach responsePeriodic wrap-up reminders
SessionEndSession closeSave session stats to database

Install Hooks

# Full setup with hooks
git clone https://github.com/rohitg00/pro-workflow.git /tmp/pw
cp -r /tmp/pw/hooks/* ~/.claude/
cp -r /tmp/pw/scripts ~/.claude/scripts/pro-workflow/
cp -r /tmp/pw/commands/* ~/.claude/commands/

Contexts & Agents

ContextWhenBehavior
devBuildingCode first, iterate
reviewPRsRead-only, security focus
researchExploringSummarize, plan
AgentPurpose
plannerBreak down complex tasks
reviewerCode review, security audit
scoutConfidence-gated exploration before implementation

Agent Teams (Experimental)

Coordinate multiple Claude Code sessions working together:

# Enable in settings.json
{ "env": { "CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS": "1" } }
  • Lead session coordinates, teammates work independently
  • Teammates message each other directly
  • Shared task list with dependency management
  • Display modes: in-process (Shift+Up/Down) or split panes (tmux/iTerm2)
  • Delegate mode (Shift+Tab): lead orchestrates only
  • Docs: https://code.claude.com/docs/agent-teams

Structure

pro-workflow/
├── .claude-plugin/
│   ├── plugin.json           # Plugin manifest
│   ├── marketplace.json      # Marketplace config
│   └── README.md
├── src/                      # TypeScript source
│   ├── db/
│   │   ├── index.ts          # Database initialization
│   │   ├── store.ts          # Stateless store factory
│   │   └── schema.sql        # SQLite schema with FTS5
│   ├── search/
│   │   └── fts.ts            # BM25 search helpers
│   └── index.ts
├── dist/                     # Compiled JavaScript
├── skills/
│   └── pro-workflow/
│       └── SKILL.md          # Main skill
├── agents/
│   ├── planner.md            # Planner agent
│   ├── reviewer.md           # Reviewer agent
│   └── scout.md              # Confidence-gated scout agent
├── commands/
│   ├── wrap-up.md
│   ├── learn-rule.md
│   ├── parallel.md
│   ├── learn.md
│   ├── search.md
│   ├── list.md
│   ├── commit.md
│   ├── insights.md
│   ├── replay.md             # Surface past learnings
│   └── handoff.md            # Session handoff document
├── hooks/
│   └── hooks.json # Hooks file
├── scripts/                  # Hook scripts
├── contexts/
│   ├── dev.md # Dev context
│   ├── review.md # Review context
│   └── research.md # Research context
├── references/
│   └── claude-code-resources.md # Claude code resources reference file
├── rules/
│   └── core-rules.md # Core rules file
├── templates/
│   └── split-claude-md/ # Split claude md template
├── package.json
├── tsconfig.json # TypeScript configuration file
└── README.md # README file

Learn Claude Code

Pro-workflow teaches Claude Code best practices directly, with links to official documentation for deep dives.

Official Docs: https://code.claude.com/docs/

Topics covered: sessions, context management, modes, CLAUDE.md, prompting, writing rules, skills, subagents, hooks, MCP, security, and IDE integration.

/pro-workflow:learn                 # Best practices guide & save learnings
/pro-workflow:learn-rule            # Capture corrections to memory
/pro-workflow:search claude-code    # Find past Claude Code learnings

SkillKit - Universal AI Skills

One skill. 32+ AI coding agents. Install pro-workflow across Claude Code, Cursor, Codex, Gemini CLI, and more with SkillKit.

# Install this skill
npx skillkit install pro-workflow

# Translate to any agent format
npx skillkit translate pro-workflow --agent cursor

# Get AI-powered skill recommendations
npx skillkit primer

Why SkillKit?

  • Install once, use everywhere
  • 15,000+ skills in the marketplace
  • Works with Claude Code, Cursor, Codex, Gemini CLI, Windsurf, and 27+ more agents

Explore the marketplace at agenstskills.com

Philosophy

  • Compound improvements - Small corrections → big gains over time
  • Trust but verify - Let AI work, review at checkpoints
  • Zero dead time - Parallel sessions keep momentum
  • Memory is precious - Both yours and Claude's

Support

If you find this useful:

  • Star this repo to help others discover it
  • Check out SkillKit for more AI coding skills
  • Report issues or suggest improvements

Distilled from Claude Code power users and real production use.

Keywords

claude-code

FAQs

Package last updated on 06 Feb 2026

Did you know?

Socket

Socket for GitHub automatically highlights issues in each pull request and monitors the health of all your open source dependencies. Discover the contents of your packages and block harmful activity before you install or update your dependencies.

Install

Related posts