Repo Year in Review

ultraworkers/claw-code

Public Rust implementation of the `claw` CLI agent harness, kept as an agent-managed 'museum exhibit'. Build-from-source only; requires an API key.

  • 195k GitHub stars
  • Rust
  • ⚖️ MIT
ultraworkers/claw-code — repo preview

What it is

Claw Code is the public Rust implementation of the claw CLI agent harness, with the canonical implementation in the rust/ workspace. The README says it is not the serious production project and is closer to a museum exhibit, an agent-managed artifact maintained by the LazyCodex and Gajae-Code harnesses. It is build-from-source only, and the crates.io claw-code crate is a deprecated stub that should not be installed.

Who it's for

  • Developers who want to inspect the Claw Code artifact and its Rust implementation of the claw CLI
  • Users willing to build a Rust CLI agent harness from source and authenticate with an API key
  • Readers following the Rust-port parity status and roadmap

Requirements

Requirements

  • Rust toolchain with cargo (Windows users install from https://rustup.rs/)
  • An API key such as ANTHROPIC_API_KEY or OPENAI_API_KEY; Claude subscription login is not a supported auth path
  • Building from source via cargo build --workspace (the README says a release build takes 5–10 minutes)

Setup

  1. Clone, build, authenticate and run

    Build the workspace from source, set your Anthropic API key, verify with claw doctor, then run a prompt or start an interactive session. Do not use cargo install claw-code; it installs a deprecated stub.

    bash
    # 1. Clone and build
    git clone https://github.com/ultraworkers/claw-code
    cd claw-code/rust
    cargo build --workspace
    
    # 2. Set your API key (Anthropic API key — not a Claude subscription)
    export ANTHROPIC_API_KEY="sk-ant-..."
    
    # 3. Verify everything is wired correctly
    ./target/debug/claw doctor
    
    # 4. Run a prompt
    ./target/debug/claw prompt "say hello"
    
    # 5. Start an interactive session
    ./target/debug/claw
  2. Windows (PowerShell) build and run

    On Windows the binary is claw.exe. After installing Rust and confirming cargo is on PATH, clone and build, then run with the .exe path.

    powershell
    $env:ANTHROPIC_API_KEY = "sk-ant-..."
    .\target\debug\claw.exe prompt "say hello"
  3. Optional: install to Cargo's bin directory

    To run claw from any directory, install from the rust/ directory into Cargo's default location (~/.cargo/bin/).

    bash
    # From the claw-code/rust/ directory
    cargo install --path . --force
    
    # Then from anywhere
    claw --help

Examples

Verify the build

bash
bash
./rust/target/debug/claw --help
./rust/target/debug/claw doctor

What it does: Run from the repo root on macOS/Linux. claw doctor is the first health check and validates your API key, model access, and tool configuration.

Run a one-shot prompt

bash
bash
./target/debug/claw prompt "say hello"

What it does: Sends a single prompt through the debug-built binary from the rust/ directory.

Credential-free PowerShell smoke test

powershell
powershell
$env:CLAW_CONFIG_HOME = Join-Path $env:TEMP "claw config home"
New-Item -ItemType Directory -Force -Path $env:CLAW_CONFIG_HOME | Out-Null
Remove-Item Env:\ANTHROPIC_API_KEY, Env:\ANTHROPIC_AUTH_TOKEN, Env:\OPENAI_API_KEY -ErrorAction SilentlyContinue
.\rust\target\debug\claw.exe help
.\rust\target\debug\claw.exe status
.\rust\target\debug\claw.exe config env
.\rust\target\debug\claw.exe doctor

What it does: Checks that the build works without live credentials by using a temporary config home and clearing API key variables.

Run the workspace test suite

bash
bash
cd rust
cargo test --workspace

What it does: Recommended after verifying the binary works.

Pros & cons

Pros

  • Pro:Detailed cross-platform build and verification guidance, including a PowerShell-first Windows path
  • Pro:claw doctor provides a built-in health check for API key, model access, and tool configuration
  • Pro:MIT-licensed, with documentation for usage, parity status, roadmap, and local OpenAI-compatible providers

Cons

  • Con:The README states it is not the serious production project and points users to LazyCodex or Gajae-Code for real work
  • Con:Build-from-source only; cargo install claw-code installs a deprecated stub that does not provide claw
  • Con:Requires an API key; Claude subscription login is unsupported, and no ACP/Zed daemon or JSON-RPC entrypoint ships yet