LaunchKit · 2026
Back to Skills

pi-workflow

Workflow orchestration for Pi's task management, self-improvement, and code quality standards.

0
319 downloads
by @kai-tw

Setup & Installation

openclaw skills install @kai-tw/pi-workflow

Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:

npx clawhub install pi-workflow

Version History

v1.1.1Feb 26, 2026 - Version note by skill developer:

Complete workflow orchestration skill with plan node, self-improvement loop, and custom hooks integration

What This Skill Does

Structured workflow system for task planning, bug fixing, and self-improvement. Covers plan-first task management, subagent delegation, and a lessons capture system with metadata that tracks recurring mistake patterns. Designed for AI agents working on multi-step or architectural tasks.

The structured lessons format with recurrence tracking and priority metadata catches repeated mistakes that plain task lists or ad-hoc notes miss.

When to use it

  • Starting a new project with a written spec before writing any code
  • Logging a correction after a mistake to prevent it from recurring
  • Fixing a reported bug autonomously without asking the user for steps
  • Tracking command failures and API errors across sessions for diagnosis
  • Running verification gates before marking a task complete

Example Workflow

Here's how your AI assistant might use this skill in practice.

INPUT

User asks: Fix the authentication bug causing 401 errors on the /api/profile endpoint

AGENT
  1. 1Enter plan mode and write a spec to tasks/todo.md listing investigation steps and acceptance criteria
  2. 2Check logs, error messages, and failing tests autonomously to identify the root cause
  3. 3Implement the minimal fix, touching only the necessary code paths
  4. 4Run tests and verify the endpoint returns 200 with valid credentials
  5. 5Update tasks/lessons.md with a structured lesson entry if a correction was made during the process
OUTPUT

Bug fixed and verified against acceptance criteria, lessons logged with Priority, Area, and Pattern-Key metadata for future reference