mistake-notebook

Record and retrieve solved mistakes in project-local and universal MISTAKE.md files.

6|Updated Feb 25, 2026
One-click install
npx skills add https://github.com/archibate/archibate-skills --skill mistake-notebook-archibate
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: mistake-notebook
Source: https://github.com/archibate/archibate-skills/tree/main/old-skills/minor-skills/mistake-notebook
Command: npx skills add https://github.com/archibate/archibate-skills --skill mistake-notebook-archibate

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Agents and developers often repeat the same severe mistakes because prior solutions are not recorded or easily discoverable; this Skill provides a simple, persistent mistake notebook to record problems, insights, and fixes for future reuse. It centralizes troubleshooting knowledge into two locations: a project-local MISTAKE.md and a universal ~/MISTAKE.md, enabling quick lookups and consistent remediation across sessions.

Core Features & Use Cases

  • Project-local vs Universal: Choose MISTAKE.md in the project root for project-scoped issues and ~/MISTAKE.md for generic programmatic mistakes that apply across projects.
  • Search and Reuse: Read existing entries to try known solutions first, reducing repeated debugging effort.
  • Append and Update Entries: When a severe problem is solved, append a structured entry with creation/update dates, project context, insights, and step-by-step solution so future agents or developers can avoid the same error.
  • Use Case: After resolving a recurring CI build failure, append the root cause and exact fix to MISTAKE.md so subsequent agents automatically check and apply that solution.

Quick Start

Ask the agent to search both MISTAKE.md and ~/MISTAKE.md for the error, apply any found solution, and if resolved append a new entry describing the problem, insights, and solution.

Frequently Asked Questions about mistake-notebook

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

FAQPage Schema
How do I prevent an AI agent from repeating the same coding mistakes in future sessions?▼

To prevent repeating coding mistakes, you can use a persistent error-tracking notebook that records solved problems, insights, and fixes into project-local or universal MISTAKE.md files. Agents query these files to apply known solutions before debugging.

What is the best way to track severe bugs and configuration errors for future reuse?▼

Tracking severe bugs and configuration errors is done by appending structured entries with dates, context, and step-by-step solutions to a MISTAKE.md file. This centralizes troubleshooting knowledge for quick lookups and consistent remediation.

How do I record postmortem entries for recurring CI build failures so they are automatically fixed?▼

To record postmortem entries for CI build failures, append the root cause and exact fix to the project root's MISTAKE.md. Subsequent agents will search this file, find the matching error, and automatically apply the documented solution.

Does this mistake tracking approach support separating project-specific issues from generic programmatic errors?▼

Yes, mistake tracking supports separating issues by using a project-local MISTAKE.md in the repository root for project-scoped problems, and a universal ~/MISTAKE.md file for generic programmatic errors that apply across all projects.

What do I need to set up before agents can start recording and retrieving solved mistakes?▼

You need to provide read and append access to MISTAKE.md files in the repository root or ~/MISTAKE.md. The agent requires the ability to search existing entries and append structured problem, insights, and solution data.

When should I not use a project-local MISTAKE.md file for agent memory?▼

You should not use a project-local MISTAKE.md for generic programmatic mistakes that apply across multiple projects. Instead, use the universal ~/MISTAKE.md file to ensure the error-tracking knowledge is available globally across sessions.