New issues and PRs from new contributors are auto-closed by default. Maintainers review auto-closed issues daily. See CONTRIBUTING.md.
Shuvpi is Shuv's maintained fork of the Shuvpi agent harness. The distinct package names, binary, config directory, and environment variables let it coexist with a native upstream Shuvpi installation.
- @shuv1337/shuvpi-coding-agent: Interactive coding agent CLI
- @shuv1337/shuvpi-agent-core: Agent runtime with tool calling and state management
- @shuv1337/shuvpi-ai: Unified multi-provider LLM API (OpenAI, Anthropic, Google, …)
The project remains derived from upstream Shuvpi. Upstream services such as the pi.dev session viewer remain external dependencies where explicitly documented.
curl -fsSL https://raw.githubusercontent.com/shuv1337/shuvpi/main/install.sh | shThe installer uses a writable npm global prefix when its bin directory is already on PATH and otherwise installs under ~/.local, without sudo or changes to npm's global configuration. Node.js 22.19.0 or newer is required.
| Package | Description |
|---|---|
| @shuv1337/shuvpi-chord | Standalone application-composition runtime for services, replicated state, RPC, and plugins |
| @shuv1337/shuvpi-telemetry | Vendor-neutral telemetry contracts, reference adapter, conformance tests, and typed schemas |
| @shuv1337/shuvpi-ai | Unified multi-provider LLM API (OpenAI, Anthropic, Google, etc.) |
| @shuv1337/shuvpi-durable | Durable conversation, task, and document runtime |
| @shuv1337/shuvpi-agent-core | Agent runtime with tool calling and state management |
| @shuv1337/shuvpi-coding-agent | Interactive coding agent CLI |
| @shuv1337/shuvpi-tui | Terminal UI library with differential rendering |
| @shuv1337/shuvpi-protocol | Shared wire protocol contracts for the server and client packages |
| @shuv1337/shuvpi-client | Client library for the Shuvpi server protocol |
| @shuv1337/shuvpi-session-backend-sqlite-node | SQLite session backend for Node.js |
| @shuv1337/shuvpi-server | Experimental Shuvpi server package |
For upstream Slack/chat automation and workflows, see earendil-works/shuvpi-chat.
Shuvpi does not include a built-in permission system for restricting filesystem, process, network, or credential access. By default, it runs with the permissions of the user and process that launched it.
If you need stronger boundaries, containerize or sandbox Shuvpi. See packages/coding-agent/docs/containerization.md for three patterns:
- Gondolin extension: keep
shuvpiand provider auth on the host while routing built-in tools and!commands into a local Linux micro-VM. - Plain Docker: run the whole
shuvpiprocess in a local container for simple isolation. - OpenShell: run the whole
shuvpiprocess in a policy-controlled sandbox.
See CONTRIBUTING.md for contribution guidelines and AGENTS.md for project-specific rules.
npm install --ignore-scripts # Install all dependencies without running lifecycle scripts
npm run build # Refresh model data, then build all packages
npm run build:offline # Rebuild using existing model data without network access
npm run check # Lint, format, and type check
./test.sh # Run tests (skips LLM-dependent tests without API keys)
./shuvpi-test.sh # Run shuvpi from sources (can be run from any directory)Shuvpi uses ~/.shuvpi and project-local .shuvpi directories. It does not automatically read, move, or delete upstream ~/.shuvpi or .shuvpi data. To reuse selected settings, copy them explicitly and review paths before launching:
mkdir -p ~/.shuvpi
cp -a ~/.shuvpi/agent ~/.shuvpi/
cp -a .shuvpi .shuvpiUse either command only when its source exists. A copy keeps the upstream installation intact.
GitHub releases include a versioned source archive covered by the release's SHA256SUMS file. Extract it and run the same build script used for the official standalone binaries:
VERSION="<release-version>"
tar -xzf "shuvpi-${VERSION}-source.tar.gz"
cd "shuvpi-${VERSION}"
./scripts/build-binaries.sh --offline-model-data --platform linux-x64 --out "$PWD/out"The archive includes release model data and native prebuilds. --offline-model-data uses that model data without refreshing provider catalogs. The script installs dependencies and builds the executable with its runtime assets; pass --skip-install if dependencies are already provided.
We treat npm dependency changes as reviewed code changes.
- Direct external dependencies are pinned to exact versions. Internal workspace packages remain version-ranged.
.npmrcsetssave-exact=trueandmin-release-age=2to avoid same-day dependency releases during npm resolution.package-lock.jsonis the dependency ground truth. Pre-commit blocks accidental lockfile commits unlessSHUVPI_ALLOW_LOCKFILE_CHANGE=1is set.npm run checkverifies pinned direct deps, native TypeScript import compatibility, and the generated coding-agent shrinkwrap.- The published CLI package includes
packages/coding-agent/npm-shrinkwrap.json, generated from the root lockfile, to pin transitive deps for npm users. - Release smoke tests use
npm run release:localto build, pack, and create isolated npm and Bun installs outside the repo before tagging a release. - Local release installs, documented npm installs, and
shuvpi update --selfuse--ignore-scriptswhere supported. - CI installs with
npm ci --ignore-scripts, and a scheduled GitHub workflow runsnpm audit --omit=devplusnpm audit signatures --omit=dev. - Shrinkwrap generation has an explicit allowlist for dependency lifecycle scripts; new lifecycle-script deps fail checks until reviewed.
If you use Shuvpi or other coding agents for open source work, please share your sessions.
Public OSS session data helps improve coding agents with real-world tasks, tool use, failures, and fixes instead of toy benchmarks.
For the full explanation, see this post on X.
To publish sessions, use badlogic/shuvpi-share-hf. Read its README.md for setup instructions. All you need is a Hugging Face account, the Hugging Face CLI, and shuvpi-share-hf.
You can also watch this video, where I show how I publish my shuvpi-mono sessions.
I regularly publish my own shuvpi-mono work sessions here:
MIT