iterate

Creates a new numbered iteration directory and updates changelog for tracked workflows.

1|Updated Jan 24, 2023
One-click install
npx skills add https://github.com/KaiNakamura/dotfiles --skill iterate-kainakamura
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: iterate
Source: https://github.com/KaiNakamura/dotfiles/tree/main/claude/skills/iterate
Command: npx skills add https://github.com/KaiNakamura/dotfiles --skill iterate-kainakamura

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

When a current approach to a task is no longer working, the iteration skill provides a structured, reproducible way to record the failure, create a clean workspace for the next attempt, and preserve context so future agents or collaborators do not repeat the same mistakes.

Core Features & Use Cases

  • Iteration discovery: Detects the current highest iteration by scanning .thoughts for iteration-NN directories.
  • Changelog updates: Appends a concise summary explaining why the iteration is being restarted into .thoughts/iterations.md to guide future agents.
  • Iteration initialization: Creates the next .thoughts/iteration-NN directory and seeds a progress.md with a high-level context summary.
  • Use Case: When an agentic workflow stalls or a prototype approach fails, invoke this flow to record why the attempt ended and begin a new iteration without altering unrelated files.

Quick Start

Start a new iteration for the current task with reason "current approach failing to converge" so the skill appends a summary to .thoughts/iterations.md, creates the next .thoughts/iteration-NN/ directory, and initializes its progress.md.

Frequently Asked Questions about iterate

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I start a new iteration when an agentic workflow stalls?

To restart an agentic workflow, invoke the iteration skill to record why the current attempt failed in .thoughts/iterations.md and create a clean .thoughts/iteration-NN directory with a seeded progress.md for the next attempt.

What is the best way to track task progress across multiple agent attempts?

Tracking task progress across agent attempts requires appending summaries to an iterations.md changelog and organizing work into sequential iteration-NN folders, ensuring future agents avoid repeating previous mistakes.

How does iteration tracking work with a .thoughts directory?

Iteration tracking scans the .thoughts directory for the highest iteration-NN folder, appends a failure summary to iterations.md, and initializes the next iteration folder with a progress.md file.

Can I restart a failed task iteration without modifying unrelated files?

Yes, restarting a failed task iteration only creates the next iteration-NN directory and updates the iterations.md changelog, leaving all unrelated files in your workspace unmodified.

When do I need to create a new iteration record for task management?

You need to create a new iteration record when your current task approach is no longer working or fails to converge, allowing you to preserve context and start a fresh attempt.

Does the iteration skill require any external dependencies to manage dotfiles?

No, the iteration skill operates without external dependencies, relying solely on the existing .thoughts directory structure to manage dotfiles and track agentic workflow progress.