openwolf-enhanced

This is an enhanced fork of OpenWolf by Cytostack Pvt Ltd. The original is a great idea; in long-lived projects its .wolf/ directory could grow without bound (multi-megabyte token ledgers, an ever-growing bug log, full-file rewrites on every edit). This fork keeps everything the original does and makes storage bounded, self-maintaining, and scopeable. The CLI is still openwolf, so it's a drop-in replacement. See what's enhanced and the CHANGELOG.

Why OpenWolf Exists

Claude Code is powerful but it works blind. It doesn't know what a file contains until it opens it. It can't tell a 50-token config from a 2,000-token module. It reads the same file multiple times in one session without noticing. It has no index of your project, no memory of your corrections, and no awareness of what it already tried.

OpenWolf gives Claude a second brain: a file index so it knows what files contain before reading them, a learning memory that accumulates your preferences and past mistakes, and a token ledger that tracks everything. All through 6 invisible hook scripts that fire on every Claude action.

What's Enhanced

Everything upstream does, plus — grouped by what it gives you:

Every change is versioned in the CHANGELOG; attribution is in the NOTICE.

Quick Start

npm install -g openwolf-enhanced

Note: this is the maintained fork. npm install -g openwolf installs the original openwolf (last released 1.0.4, March 2026, unmaintained) — a different package. Install openwolf-enhanced for the bounded-storage, self-maintenance and security work described above. Both provide the same openwolf command.

git clone https://github.com/bassprofressor-lab/openwolf-enhanced.git
cd openwolf-enhanced
pnpm install
pnpm build            # builds CLI, hooks, and dashboard
npm install -g .      # installs the `openwolf` command globally

Then, in any project:

cd your-project
openwolf init

That's it. Use claude normally. OpenWolf is watching.

Rebuilding an installed copy

Two things to know before you re-run pnpm build on a copy you have already installed globally:

• prebuild deletes dist/ before compiling, and the global openwolf command is a symlink into dist/bin/openwolf.js. If the build then fails, the CLI is gone until you build again. Back dist/ up first if you are mid-change, or build in a clean checkout.
• A rebuild does not deploy anything. The hooks that actually run are per-project copies under <project>/.wolf/hooks/, invoked by .claude/settings.json — not the installed package. Run openwolf update afterwards to push new hooks into your projects (it touches every registered project; scope it with --project <name>).

Upgrading the tool? The hooks that actually run are per-project copies in <project>/.wolf/hooks/, not the global package. After rebuilding/reinstalling, run openwolf update (or openwolf update --project <name>) to copy the new hooks into your projects — a global reinstall alone does not update them. openwolf update with no --project updates all registered projects.

What It Creates

openwolf init creates a .wolf/ directory in your project:

How It Works

Before Claude reads a file, OpenWolf tells it what the file contains and how large it is. If Claude already read that file this session, OpenWolf warns it. Before Claude writes code, OpenWolf checks your cerebrum.md for known mistakes. After every write, it auto-updates the project map and logs token usage. You see none of this. It just happens.

You type a message
    ↓
Claude decides to read a file
    ↓
OpenWolf: "anatomy.md says this file is ~380 tokens. Description: Main entry point."
    ↓
Claude reads the file → OpenWolf logs the read, checks for repeated reads
    ↓
Claude writes code → OpenWolf checks cerebrum.md for known mistakes
    ↓
Claude finishes → OpenWolf updates anatomy.md, appends to memory.md, updates the ledger

Keeping .wolf/ Healthy

The .wolf/ directory is designed to stay small, but on very active projects you can compact it any time — no daemon required:

openwolf doctor --dry-run   # report footprint + warnings, change nothing
openwolf doctor             # compact ledger, consolidate memory, dedup buglog,
                            # prune backups, rotate logs, clear tmp

openwolf status shows the current footprint and warns before anything gets large.

Tuning limits

Edit the openwolf.retention block in .wolf/config.json (defaults shown):

{
  "openwolf": {
    "retention": {
      "token_ledger_max_sessions": 200,
      "session_io_max": 100,
      "buglog_max_entries": 200,
      "backups_keep": 10,
      "memory_consolidate_after_days": 7,
      "memory_max_bytes": 262144,
      "daemon_log_max_bytes": 524288
    }
  }
}

These survive openwolf update (config is deep-merged, not overwritten).

Scoping with .wolfignore

Create a .wolfignore at your project root to exclude paths from anatomy scanning and hook tracking (gitignore-style):

vendor/
dist/
**/*.generated.ts
*.log

Commands

openwolf init                 Initialize .wolf/ and register hooks
openwolf status               Show health, stats, .wolf/ footprint, size warnings
openwolf doctor               Report + compact .wolf/, suggest .wolfignore [--dry-run]
openwolf recall <query>       Keyword-search .wolf + Claude's native memory [--limit N] [--json]
openwolf export <what>        Export