Skip to content

Musts

AI Agents & Assistants

A framework that enforces local, code‑adjacent validation checks so AI agents can only claim work is done when all declared tests, builds, and rules have fresh evidence.

Track releases GitHub
Rust Latest musts-v0.1.9 · 22d ago Security brief →

Features

  • Declarative `MUSTS.yml` manifests define validation checks next to the code they protect
  • Generates concrete pending tasks with exact shell commands via `musts validate`
  • Records and verifies evidence (logs, screenshots, JSON reports) using `musts evidence`
  • Supports multiple built‑in capabilities (cargo test/format/clippy, Bazel build, custom agent judgements)
  • Works with any AI code assistant that can run shell commands

Recent releases

View all 22 releases →
No immediate action
musts-v0.1.9 New feature

musts run + tighten evidence

Open
No immediate action
musts-core-v0.1.7 Breaking risk

Add musts run

Open
No immediate action
musts-protocol-v0.1.1 New feature

`musts run` + tighter evidence

Open
No immediate action
musts-v0.1.8 New feature

Validation before commit

Open
No immediate action
musts-core-v0.1.6 New feature

Pre-commit validation + exclude_paths

Open

Weekly OSS security release digest.

The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.

No spam, unsubscribe anytime.

About

Stars
7
Forks
1
Languages
Rust Shell Makefile
View on GitHub

Install & Platforms

Install via
brew cargo binary

Similar tools

Give your AI agent a brain that understands your codebase
Termaxa
Ait
Circus Chief
Goose

Beta — feedback welcome: [email protected]