ClaudeForge is a free, open-source toolkit that analyzes your codebase and generates or maintains a CLAUDE.md for you, with a background agent that keeps it in sync. It is a genuinely useful maintenance layer for an existing project. Brainfile solves a different first problem: what do you write in a CLAUDE.md before you have one, for a role ClaudeForge's codebase-scanner was never built to understand — founder, marketer, trader, consultant, and eleven more.
ClaudeForge: a free, MIT-licensed CLAUDE.md toolkit for Claude Code: 5 skills, 3 slash commands (/enhance-claude-md, /sync-claude-md, /claude-to-agents) and a background 'claude-md-guardian' agent that scores file quality 0-100 and enforces a hard 150-line cap.
Brainfile: a library of maintained CLAUDE.md configurations — structured role and project context, pre-written for your profession, that you customize and Claude Code reads automatically from your project root.
These are not the same starting point. ClaudeForge analyzes code that already exists and audits or regenerates the file from it. Brainfile hands you a written file when there is no codebase to analyze yet, or when the context you need is about your role, not your repo.
| Feature | ClaudeForge | Brainfile |
|---|---|---|
| How it gets written | Analyzes your existing codebase and generates from it | Pre-authored per role, you fill in the specifics |
| Works before code exists | No — needs a repo to scan | Yes — role and project context, not a code scan |
| Non-developer roles | Built for developer/repo workflows | Developer, founder, trader, marketer and 11 more |
| Length constraint | Hard 150-line cap enforced at load/write time | No fixed cap — sized to what the role actually needs |
| Ongoing maintenance | Guardian agent + weekly /sync-claude-md you run yourself | Configurations updated as Claude's behavior changes |
| Setup requirement | Claude Code 2.1.4+ and git, plus installing the toolkit | None beyond the Claude subscription you already have |
| Cost | Free and open source (MIT) | Paid subscription — see the pricing page for current plans |
| Transparency | Open source, you can read the hooks and skills | Plain markdown you read, edit and version-control |
You already have a codebase and want a free, open-source tool that scans it, drafts a CLAUDE.md from what it finds, and keeps nudging you to re-sync as the code moves. That is exactly what it is built for.
You want a written starting point for a role ClaudeForge's code-scanner does not model — or you want one before there is a codebase to scan at all. You want something to customize in 30 minutes, not a tool to install, run and review.
This is a reasonable pairing, not a contradiction: start a repo from a Brainfile configuration for the role and conventions, then let ClaudeForge's guardian agent watch that file for drift as the code changes. One authors the starting judgment; the other audits it stays in sync.
ClaudeForge's model is analysis-first: point it at a repository, it explores the code and drafts a CLAUDE.md from what it observes, then a background agent keeps checking that file against the code going forward. That is real, useful engineering for a tool built to watch a codebase.
Brainfile's model is authorship-first: someone already thought through what a founder, a trader, or a developer needs Claude to know, and wrote it down. You start from that judgment and edit it — there is no scan step, and no codebase is required for it to make sense.
15 production-ready CLAUDE.md configurations, updated as Claude evolves. Works with the Claude subscription you already have.
30-day money-back guarantee · Instant access · Cancel anytime
AI context engineering tips and new configuration releases. Free.
No spam. Unsubscribe anytime.