Forge is an open-source Claude Code plugin (MIT) that adds governance to multi-agent AI coding workflows. 22 specialized agents, file locking, drift detection, knowledge capture, automated quality gates.

4,590 tests. Rust orchestrator + React dashboard + Claude Code plugin system.

Demo: https://youtu.be/4yRYimZYzBw Site: https://forge.nxtg.ai

Comments URL: https://news.ycombinator.com/item?id=47414331

Points: 2

# Comments: 1

Which repos have zero tests? - Are my React versions consistent across 3 frontends? - Which project hasn't been committed to in weeks? - Where am I duplicating FastAPI patterns instead of sharing a library?

I got tired of manually checking, so I built Atlas. It's a Python CLI that scans your repos locally and gives you a portfolio-level health dashboard.

What it does Point it at your projects, run atlas scan, and you get this:

  pip install nxtg-atlas

  atlas init --name "My Portfolio"
  atlas add ~/projects/api
  atlas add ~/projects/frontend
  atlas scan
  atlas status

+---------+--------------------+----------+--------+----------+---------------------+ | | Project | Health | Tests | LOC | Tech Stack | +---------+--------------------+----------+--------+----------+---------------------+ | (green) | Dx3 | A 94% | 179 | 164,806 | Python . TypeScript | | (green) | content-engine | B+ 88% | 23 | 12,552 | Python | | (green) | Podcast-Pipeline | B+ 88% | 49 | 33,680 | Python . FastAPI | | (green) | nxtg.ai | B+ 87% | 36 | 49,938 | TypeScript . Next.js| | (green) | Faultline | B+ 86% | 29 | 15,527 | TypeScript . React | | (green) | voice-jib-jab | B+ 86% | 58 | 45,311 | TypeScript . Python | | (yellow)| SynApps | C 67% | 114 | 793,406 | Python . TypeScript | | (red) | NXTG-Forge | D 59% | 121 | 136,083 | TypeScript . Rust | +---------+--------------------+----------+--------+----------+---------------------+

8 Projects | 609 Test Files | 1,251,303 LOC | Scanned in 31 seconds That's not a mockup. That's the actual output from scanning our production portfolio. The C and D grades were embarrassing -- but that's the point. You can't fix what you can't see.

How health scoring works Each project gets an A-F grade based on 4 weighted dimensions:

- Tests (35%) -- test file count relative to source files - Structure (25%) -- CI/CD, .gitignore, package config, linting, source organization - Git Hygiene (20%) -- commit history, remote status, clean working tree - Documentation (20%) -- README, CHANGELOG, docs/, etc.

Cross-project intelligence This is where it gets interesting. Atlas compares deps and versions across all your repos:

- Shared dependencies (fastapi used in 4 projects, react in 3) - Version mismatches (react ^18.2.0 in one app, ^19.2.1 in another) - Health gaps (3 projects with zero tests, 1 with 50+ uncommitted changes) - These are the things that bite you six months from now.

What it is NOT - Not a cloud service. Zero network calls, no telemetry, no accounts. - Not a linter or code quality tool. It doesn't read your source code line by line. - Not a replacement for SonarQube / CodeClimate. Those go deep on one repo. Atlas goes wide across many.

Details - Python 3.11+, built with Typer + Rich - State is a single JSON file at ~/.atlas/portfolio.json - Detects 10+ languages, 15+ frameworks, 5+ databases - MIT license, 100% open source — no tiers, no limits, no feature gates - Everything works: unlimited projects, cross-project intelligence, export, batch-add

--> GitHub: https://github.com/nxtg-ai/repoatlas --> PyPI: https://pypi.org/project/nxtg-atlas/

If atlas saves you time, consider supporting development.

Happy to answer questions about the scoring algorithm, the detection heuristics, or how we're using it internally.

Comments URL: https://news.ycombinator.com/item?id=47242815

Points: 1

# Comments: 0

I ran into repeated issues with agents editing the same files, losing decisions, and drifting away from the spec. I ended up building a small Rust MCP server that adds task claiming, file-level locking, and drift checks against SPEC.md.

It’s intentionally boring infrastructure… single binary, file-based state, deterministic orchestration. The AI tools stay interchangeable.

Repo if anyone’s curious: https://github.com/nxtg-ai/forge-orchestrator

Forge is an orchestration layer that sits between AI coding tools and your codebase. It's a single Rust binary (~3 MB, zero runtime deps) that runs as an MCP server over stdio. Any MCP-compatible AI tool can call it.

What it does: - File locking: When an agent claims a task, Forge locks the target files. Other agents see the lock and work on something else. Conflicts become structurally impossible. - Knowledge flywheel: Agents call forge_capture_knowledge to store decisions, patterns, and gotchas. Other agents query the store before making decisions. Knowledge compounds across sessions instead of evaporating. - Drift detection: Sends recent changes + the project spec to an LLM for alignment scoring. Catches "you were supposed to build auth but you're refactoring CSS" before it compounds. - Governance: 5-dimension health check (tests, security, docs, architecture, git hygiene) that agents and humans can query at any time.

The brain is pluggable: a free heuristic engine (pattern matching, works offline) or an LLM engine (GPT-4.1 by default). Switch with one CLI command.

State is a single JSON file in .forge/ — human-readable, git-trackable, zero operational overhead.

51 tests (30 unit, 9 CLI, 12 MCP protocol), 0 compiler warnings, 0 unsafe blocks.

MIT licensed. Whitepaper with the full architecture: https://nxtg.ai/insights/forge-whitepaper

Happy to answer questions about the Rust implementation, MCP protocol design, or the multi-agent coordination problem in general.

Comments URL: https://news.ycombinator.com/item?id=46943041

Points: 1

# Comments: 1

Forge is an orchestration layer that sits between AI coding tools and your codebase. It's a single Rust binary (~3 MB, zero runtime deps) that runs as an MCP server over stdio. Any MCP-compatible AI tool can call it.

MIT licensed. Whitepaper with the full architecture: https://nxtg.ai/insights/forge-whitepaper