How does Compound Docs detect when a problem is solved?

The skill monitors for confirmation phrases like 'that worked', 'it's fixed', or 'problem solved', and can also be triggered manually using the /doc-fix or /compound commands.

Where are the documentation files stored?

Files are stored within your repository in the .ai/docs/solutions/ directory, organized into subfolders based on the specific category of the problem.

What happens if a similar problem has already been documented?

During the capture process, the skill searches your existing documentation. If a match is found, it prompts you to either create a cross-reference, update the existing doc, or link them as related issues.

What kind of metadata does it capture?

It uses a validated YAML frontmatter to track module names, symptoms, root causes, investigation steps, environment details (like Rails version), and problem severity.

Can I promote a specific fix to a 'Critical Pattern'?

Yes. After documentation, the skill provides a decision menu where you can choose to promote the solution to a critical pattern file that AI agents are required to read before generating code.

Compound Docs

Name: Compound Docs
Author: SiTaggart

bySiTaggart

学習とドキュメンテーション

Automatically captures and organizes solved technical problems into a searchable, schema-validated knowledge base.

概要

Compound Docs is a specialized Claude Code skill designed to build and maintain institutional knowledge by documenting solved problems in real-time. It monitors conversations for solution confirmations and triggers a structured 7-step workflow to extract context, validate metadata via YAML frontmatter, and categorize documentation for fast retrieval. This skill prevents the repetition of complex debugging efforts by transforming ephemeral chat fixes into a permanent, organized directory of 'solutions' that future AI sessions can reference and learn from.

主な機能

Automated trigger detection from conversational success phrases
Automatic directory organization based on problem-type mapping
0 GitHub stars
Cross-referencing logic to link related issues and prevent duplication
Strict YAML schema validation for consistent metadata and searchability
Decision menu for promoting critical patterns to 'Required Reading'

ユースケース

Onboarding new team members or AI agents with a 'Required Reading' pattern library
Standardizing documentation across complex, multi-module Rails or AI-driven projects
Building a searchable technical repository of non-obvious bug fixes

概要

主な機能

Automated trigger detection from conversational success phrases
Automatic directory organization based on problem-type mapping
0 GitHub stars
Cross-referencing logic to link related issues and prevent duplication
Strict YAML schema validation for consistent metadata and searchability
Decision menu for promoting critical patterns to 'Required Reading'

ユースケース

Onboarding new team members or AI agents with a 'Required Reading' pattern library
Standardizing documentation across complex, multi-module Rails or AI-driven projects
Building a searchable technical repository of non-obvious bug fixes