How does Compound Docs detect when a problem is solved?

The skill monitors the conversation for confirmation phrases like 'that worked' or 'it's fixed,' and can also be manually invoked using the /doc-fix command.

Does it document every minor fix or typo?

No, the skill is configured to skip trivial syntax errors and focus on non-obvious solutions or complex debugging efforts that provide long-term value.

Where are the documented solutions stored?

Solutions are stored as Markdown files within a structured directory at docs/solutions/, organized into subcategories based on the problem type.

What is the 'Required Reading' feature?

It allows users to promote specific critical solutions to a pattern file that Claude reviews before generating new code, preventing the repetition of foundational mistakes.

Can I link related issues together?

Yes, the skill includes a cross-referencing step that searches for similar historical issues and adds links between the new documentation and existing records.

Compound Documentation Manager

Name: Compound Documentation Manager
Author: jjeremycai

byjjeremycai

0•

Learning & Documentation

Automates the capture and categorization of solved coding problems into a searchable institutional knowledge base.

Compound Docs is a specialized skill designed to transform transient troubleshooting sessions into permanent institutional knowledge. It automatically detects when a technical problem has been resolved, gathers critical context—including root causes, environment details, and specific code fixes—and generates structured Markdown documentation. By utilizing validated YAML frontmatter and standardized categorization, it builds a searchable repository that prevents the recurrence of past bugs and allows users to promote non-obvious solutions to 'Required Reading' for future AI sessions.

Key Features

01Automatic trigger detection for solved problems and verification phrases

02Standardized categorization and file naming for organized knowledge storage

03Post-capture decision menu for promoting solutions to critical patterns

04Cross-referencing logic to link related historical issues and patterns

05Validated YAML frontmatter for consistent metadata and high searchability

060 GitHub stars

Use Cases

01Creating 'Required Reading' lists to improve AI accuracy in future coding sessions

02Building a searchable technical debt and troubleshooting repository for complex modules

03Standardizing documentation for non-obvious bug fixes across a development team

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jjeremycai/claudekit compound-docs

For use in Claude.ai and ChatGPT

Download Skill