SCALE Engine CLI
SCALE Engine constrains agents: quality gates, test evidence, hooks against dangerous commands and an evidence log.
- 59 ★ on the repository
- Last commit
Extended access
Why this level
- The installer writes rule files and hooks into the project, starts a local panel daemon and can create a Git repository, and the hooks block agent commands.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add cli/scale-engineDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The installer writes rule files and hooks into the project, starts a local panel daemon and can create a Git repository, and the hooks block agent commands.
This is third-party code. Review the repository files before installing.
What it does
SCALE Engine is an engineering scaffold engine for agents: Claude Code, Codex CLI, OpenCode, Cursor, Gemini CLI and others. The idea is to constrain the agent physically rather than by a prompt request. One command, `npx -y @hongmaple0820/scale-engine@latest install --dir .`, detects the project, prepares Git, creates the workflow, writes rules into `.scale/` and shows an acceptance summary. It includes a stage machine from definition to ship, gates for build, lint, tests, security and review, a record of real commands and verification results in an evidence store, and control of commits and overlaps between sessions. The Shield compiles YAML policies into hooks that block more than forty dangerous commands, and the control panel runs at `http://127.0.0.1:3210`.
Who it is for. Teams building real projects with several agents who are tired of unproven claims that tests passed and of chaos in Git.
Good fit when
- An agent says tests passed and you need to record which command and result confirmed it.
- Several agent sessions edit the same files and overlaps must be visible in advance.
- Dangerous commands must be blocked at hook level rather than requested in a prompt.
- A multi-repository project needs coordinated branches, merges and shipping.
Not a fit when
- A small personal project: the `.scale/` folder and quality gates would outweigh the task.
- English documentation is needed: the main README is in Chinese.
- Rule files, reports and a manifest must not be added to the repository.
Example request
Install SCALE Engine in this project, run a preflight check and show which quality gates are enabled.Limitations
The main documentation is in Chinese. Git 2.28 or newer is needed for Git governance. Installation adds a `.scale/` folder with rules, reports and a manifest to the project, and for new repositories commits only `.gitignore` and explicitly created configs. The licence is not defined in the repository metadata, so read the terms in the License section. For CI environments the `--no-git` flag skips Git operations. The project has many features and a full description goes beyond this card.
How to disable. Run `scale uninstall`: evidence is archived before removal. Hooks and rule files can also be deleted by hand, and `scale clean` only previews what would be removed by default.
FAQ
What gets created in the project?
A `.scale/` folder with verification profiles, skill routing, tool rules, report templates and a manifest, plus standards in `docs/standards/`.
Which agents are supported?
Claude Code, Codex CLI, OpenCode, Cursor, Gemini CLI and others: the README states support for 23 agents and platforms.
How do I see the state?
`scale open --dir .` opens the panel at `http://127.0.0.1:3210` and `scale doctor surface` shows everything SCALE wrote into the project.
What does the Shield do?
It compiles YAML policies into hook scripts and blocks more than forty dangerous commands, including native PowerShell and cmd forms.
How do I remove SCALE?
With `scale uninstall`, which archives evidence before removal. `scale clean` previews and by default deletes nothing.
Discussion
No questions or reviews yet.
Related
CLI toolsA 13-skill delivery workflow for coding agents: from idea to plan, review, security audit, tests and a shipped pull request
A skill that runs an eight-step loop: Claude orchestrates and arbitrates while Codex writes, critiques and fixes the code.
Shared Claude Code settings for a team: hooks against dangerous commands, a check before finishing and skills from spec to review
Codely team skills that run agent work in phases: a plan in a file or GitHub issues, review after each phase and a clean commit
A Claude Code plugin that limits the agent with rules and hooks and hands independent code review to Codex CLI in a review loop.
A CLI that generates and maintains AI context files, CLAUDE.md, Cursor rules, AGENTS.md and Copilot instructions, keeping them accurate as code changes
README in short
Covers one-command installation, a short agent workflow, the problems the engine addresses, a three-minute quick start, core capabilities and the three-engine architecture.