How does the cache-first strategy work?

The skill first checks the local Codex cache for the requested document. If found and not expired, it returns the content instantly. If missing or expired, it fetches the document from the source and updates the cache.

Does this skill require the Fractary CLI to be installed?

Yes, this skill delegates all retrieval operations to the Fractary CLI. If the CLI is missing, the skill will provide structured installation instructions.

Can I force Claude to fetch the latest version of a document?

Yes, you can use the bypass_cache parameter or set a custom TTL (Time-To-Live) to ensure Claude retrieves the freshest version from the source repository.

What is the codex:// URI format used by this skill?

The skill uses a standardized format: codex://{org}/{project}/{path}. For example: codex://fractary/auth-service/docs/oauth.md.

Is document access secure?

Access is managed by the Fractary CLI and SDK, which perform permission checks based on document frontmatter and your local configuration before returning content.

Codex Document Fetcher

Name: Codex Document Fetcher
Author: fractary

$fractary$ byfractary

•

Aprendizaje y Documentación

Retrieves documents from the Fractary codex knowledge base using a high-performance, cache-first retrieval strategy.

The Codex Document Fetcher skill enables Claude to seamlessly access and integrate domain-specific documentation directly into development workflows by delegating to the Fractary CLI. It provides a secure, cache-first mechanism for fetching content via standardized codex:// URIs, handling complex tasks like permission checking, multi-source retrieval (GitHub, S3, HTTP), and automatic cache management. This ensures Claude always has the most relevant and up-to-date context from your organizational knowledge base without requiring manual file uploads or context-heavy prompts.

Características Principales

01Configurable TTL and cache-bypass options for real-time document updates

022 GitHub stars

03Built-in permission validation based on document frontmatter

04Cache-first document retrieval for sub-100ms response times on hits

05Standardized codex:// URI support for consistent cross-project referencing

06Multi-source support including GitHub, S3, and HTTP via Fractary SDK

Casos de Uso

01Retrieving internal security policies or coding standards for automated compliance checks

02Accessing project-specific documentation stored across decentralized cloud repositories

03Fetching architectural guidelines or API specifications during code generation tasks

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add fractary/claude-plugins document-fetcher

For use in Claude.ai and ChatGPT

Download Skill