fwdslsh

CLI for sharing extensible assets across OpenCode and Claude Code platforms. Don't silo your workflows.

$ hyphn init
$ hyphn setup
$ hyphn sources add ~/team-hyphn

Why we built it: AI coding tools shouldn't silo your custom workflows. Share them.

Simple convention for publishing content that both humans and AI can reliably navigate using .burrow.json manifests.

$ cat .burrow.json
$ rabit validate
$ rabit fetch

Why we built it: AI agents need structured paths, not HTML soup.

Containerized dev environment for safely running Claude AI and CLI agents without risking your host system.

$ dispatch init
$ dispatch start
$ dispatch attach

Why we built it: Run AI assistants in complete isolation with enterprise-grade security.

CLI tool for retrieving and bundling relevant context from the Hyph3n learning system with token budget management.

$ disclose learnings --query "auth"
$ disclose session-context
$ disclose pack

Why we built it: AI needs the right context, not everything.

Crawl websites, download Git repos, ingest feeds - all converted to clean Markdown.

$ gather https://docs.example.com
$ gather github.com/owner/repo/tree/main/docs

Why we built it: Documentation migration shouldn't require frameworks.

Crawl websites, extract main content, and convert to clean Markdown. Powered by Bun for maximum performance with zero dependencies.

$ inform https://docs.example.com
$ inform github.com/owner/repo/tree/main/docs
$ inform --include "*.md" --exclude "temp/*"

Why we built it: Documentation migration and content extraction shouldn't require heavy frameworks. Fast, focused, effective.

Generate llms.txt and llms-full.txt from Markdown/HTML directories. Perfect for AI-powered documentation workflows and searchable knowledge bases.

$ catalog --input docs --output build
$ catalog --sitemap --base-url https://docs.example.com
$ catalog --optional "drafts/**/*" --validate

Why we built it: AI tools need structured, indexed content. We made it trivial to prepare documentation for AI consumption.