# claw-code-agent
**Repository Path**: quickn/claw-code-agent
## Basic Information
- **Project Name**: claw-code-agent
- **Description**: No description available
- **Primary Language**: Python
- **License**: Not specified
- **Default Branch**: main
- **Homepage**: None
- **GVP Project**: No
## Statistics
- **Stars**: 0
- **Forks**: 0
- **Created**: 2026-06-09
- **Last Updated**: 2026-06-09
## Categories & Tags
**Categories**: Uncategorized
**Tags**: None
## README
Claw Code Agent
A Python reimplementation of the Claude Code agent architecture โ local models, full control, zero dependencies.
---
## ๐ข What's New
> **April 2026 โ Major Update**
| | Feature | Details |
|---|---------|---------|
| ๐ | **Interactive Chat Mode** | New `agent-chat` command โ multi-turn REPL with `/exit` to quit |
| ๐ | **Streaming Output** | Token-by-token streaming with `--stream` flag |
| ๐ | **Plugin Runtime** | Full manifest-based plugin system โ hooks, tool aliases, virtual tools, tool blocking |
| ๐ | **Nested Agent Delegation** | Delegate subtasks to child agents with dependency-aware topological batching |
| ๐ | **Agent Manager** | Lineage tracking, group membership, batch summaries for nested agents |
| ๐ | **Custom Agent Profiles** | Discover local markdown-defined agents from `~/.claude/agents` and `./.claude/agents` and use them through the `Agent` tool |
| ๐ | **Cost Tracking & Budgets** | Token budgets, cost budgets, tool-call limits, model-call limits, session-turn limits |
| ๐ | **Structured Output** | JSON schema response mode with `--response-schema-file` |
| ๐ | **Context Compaction** | Auto-snip, auto-compact, and reactive compaction on prompt-too-long errors |
| ๐ | **File History Replay** | Journaling of file edits with snapshot IDs, replay summaries on session resume |
| ๐ | **Truncation Continuation** | Automatic continuation when model response is cut off (`finish_reason=length`) |
| ๐ | **Ollama Support** | Works out of the box with Ollama's OpenAI-compatible API |
| ๐ | **LiteLLM Proxy Support** | Route through LiteLLM Proxy to any provider |
| ๐ | **OpenRouter Support** | Cloud API gateway โ access OpenAI, Anthropic, Google models via one endpoint |
| ๐ | **Query Engine** | Runtime event counters, transcript summaries, orchestration reports |
| ๐ | **Remote Runtime** | Manifest-backed local remote profiles, connect/disconnect state, and remote CLI/slash flows |
| ๐ | **Hook & Policy Runtime** | Local `.claw-policy.json` / hook manifests with trust reporting, safe env, tool blocking, and budget overrides |
| ๐ | **Task & Plan Runtime** | Persistent local tasks and plans with plan-to-task sync and dependency-aware task execution |
| ๐ | **MCP Transport** | Real stdio MCP transport for `initialize`, resource listing/reading, and tool listing/calling |
| ๐ | **Search Runtime** | Provider-backed `web_search` with local manifests, activation state, and `/search` flows |
| ๐ | **Config & Account Runtime** | Local config/settings mutation plus manifest-backed account profiles and login/logout state |
| ๐ | **Ask-User Runtime** | Queued or interactive local ask-user flow with history, slash commands, and agent tool support |
| ๐ | **Team Runtime** | Persisted local teams and message history with team/message tools and slash/CLI inspection |
| ๐ | **Notebook Edit Tool** | Native `.ipynb` cell editing through the real agent tool registry |
| ๐ | **Workflow Runtime** | Manifest-backed local workflows with workflow tools, slash commands, and run history |
| ๐ | **Remote Trigger Runtime** | Local remote triggers with create/update/run flows similar to the npm remote trigger surface |
| ๐ | **Worktree Runtime** | Managed git worktrees with mid-session cwd switching, slash commands, and CLI flows |
| ๐ | **Tokenizer-Aware Context** | Cached tokenizer backends with heuristic fallback for `/context`, `/status`, and compaction |
| ๐ | **Prompt Budget Preflight** | Preflight prompt-length validation, token-budget reporting, and auto-compact/context collapse before backend failures |
| ๐ | **LSP Runtime** | Local LSP-style code intelligence for definitions, references, hover, symbols, call hierarchy, and diagnostics |
| ๐ | **Local Web GUI** | Browser-based chat UI via `python -m src.gui` โ modern dark theme, slash command palette, session browser, settings panel |
| ๐ | **Pasted-Content Refs** | Pastes โฅ500 chars into the GUI composer collapse to `[Pasted text #N +M lines]` chips and re-expand server-side before the agent runs |
| ๐ | **GUI Runtime Knobs** | Settings panel exposes temperature, per-turn timeout, streaming toggle, and max-turns โ all round-tripped live through `/api/state` |
| ๐ | **GUI Budgets & Limits** | Advanced settings disclosure for every `BudgetConfig` field: cost ceiling, token budgets, tool/model call caps, delegated task cap, session turn cap โ blank input clears the limit |
| ๐ | **GUI System Prompt & Schema** | Custom / append / override system prompts and a structured-output JSON schema editor (with strict toggle) live-editable in the settings panel |
| ๐ | **GUI Context Management** | Auto-snip / auto-compact thresholds, compact-preserve count, CLAUDE.md discovery toggle, and additional working directories โ all editable from the settings panel and the new `--auto-snip-threshold` / `--auto-compact-threshold` / `--add-dir` flags |
| ๐ | **GUI Tasks View** | Browse, create, start, complete, and cancel local tasks from a new **Tasks** tab; mutations call straight into `TaskRuntime` so completing a task auto-unblocks dependents just like the slash-command path |
| ๐ | **GUI Plan View** | Edit the local porting plan (steps + explanation + per-step status/priority) from a new **Plan** tab; saves go through `PlanRuntime.update_plan` and optionally sync to the task list |
| ๐ | **GUI Memory View** | Browse, edit, create, and delete the discovered `CLAUDE.md` / `.claude/rules/*.md` memory files from a new **Memory** tab; writes are sandboxed to the workspace + `~/.claude` |
| ๐ | **GUI File History View** | New **History** tab aggregates `file_history` entries from every saved session (newest first) โ one row per shell run / file edit / nested agent call with snapshot ids and changed paths |
| ๐ | **GUI Background Sessions** | New **Background** tab lists detached `agent-bg` runs (running/exited/completed/failed), shows live logs, and lets you kill a running session โ same `BackgroundSessionRuntime` the CLI uses |
| ๐ | **GUI Worktree View** | New **Worktree** tab โ show status & history, create a managed `git worktree` (auto-switches the agent's cwd), and exit it (keep or remove); state survives reload via `WorktreeRuntime` |
| ๐ | **GUI Skills Marketplace** | New **Skills** tab โ card grid of every bundled skill with description, when-to-use, aliases, and allowed tools; "Use in chat" button drops the invocation into the composer |
| ๐ | **GUI Accounts View** | New **Accounts** tab โ discover profiles from `.claude/account.json`, log in by name or with an ephemeral identity, view login/logout history; persists into `AccountRuntime` state |
| ๐ | **GUI Remote Profiles** | New **Remote** tab โ discover remote/SSH/teleport/direct-connect/deep-link profiles from `.claw-remote.json` etc., connect by name or ephemeral target, view connect/disconnect history |
| ๐ | **GUI MCP Servers** | New **MCP** tab โ list discovered servers/resources/tools from `.claw-mcp.json`/`.mcp.json`, read inline + stdio resources, call tools with custom JSON args; "Probe stdio servers" toggle controls subprocess cost |
| ๐ | **GUI Plugins View** | New **Plugins** tab โ list manifests from `.claw-plugin/plugin.json`, `.codex-plugin/plugin.json`, and `plugins/*/plugin.json` with their tools, virtual tools, aliases, blocks, and lifecycle hooks |
| ๐ | **GUI Ask-User Queue** | New **Ask** tab โ preload answers (exact or contains match), browse the queue and history, and clear past entries; the agent's `Ask` tool consumes them straight from `.port_sessions/ask_user_runtime.json` |
| ๐ | **GUI Workflows View** | New **Workflows** tab โ list discovered workflow definitions from `.claw-workflows.json`, trigger a recorded run with custom JSON arguments, browse run history |
| ๐ | **GUI Search View** | New **Search** tab โ discover providers from `.claw-search.json`/`.claude/search.json`, activate one, and run live SearXNG/Brave/Tavily queries straight from the browser |
| ๐ | **GUI Remote Triggers** | New **Triggers** tab โ list/create/run remote triggers (manifest-defined or local), record run history; mirrors `RemoteTriggerRuntime` exactly |
| ๐ | **GUI Teams View** | New **Teams** tab โ create teams with members, send messages between them, view full message history; persisted via `TeamRuntime` |
| ๐ | **GUI Diagnostics Tab** | New **Diag** tab โ render the existing markdown reports (`summary`, `manifest`, `parity-audit`, `setup-report`, `command-graph`, `tool-pool`, `bootstrap-graph`) on demand without shelling out |
| ๐ | **Daemon Commands** | Local `daemon start/ps/logs/attach/kill` wrapper over background agent sessions |
| ๐ | **Background Sessions** | Local `agent-bg`, `agent-ps`, `agent-logs`, `agent-attach`, and `agent-kill` flows |
| ๐ | **Testing Guide** | Comprehensive [TESTING_GUIDE.md](TESTING_GUIDE.md) with commands for every feature |
| ๐ | **Parity Checklist** | Full [PARITY_CHECKLIST.md](PARITY_CHECKLIST.md) tracking implementation status vs npm source |
---
## ๐ About
This repository reimplements the [Claude Code](https://docs.anthropic.com/en/docs/claude-code) npm agent architecture **entirely in Python**, designed to run with **local open-source models** via an OpenAI-compatible API server.
Built on the public porting workspace from [instructkr/claw-code](https://github.com/instructkr/claw-code), the active development lives at [HarnessLab/claw-code-agent](https://github.com/HarnessLab/claw-code-agent).
> **Goal:** Not to ship the original npm source, but to reimplement the full agent flow in Python โ prompt assembly, context building, slash commands, tool calling, session persistence, and local model execution.
>
> **Zero external dependencies** โ just Python's standard library.
---
## โจ Key Features
| Feature | Description |
|---------|-------------|
| ๐ค **Agent Loop** | Full agentic coding loop with tool calling and iterative reasoning |
| ๐ฌ **Interactive Chat** | Multi-turn REPL via `agent-chat` with session continuity |
| ๐ฅ๏ธ **Local Web GUI** | Browser-based chat UI launched with `python -m src.gui` โ sessions browser, slash command palette, live settings |
| ๐งฐ **Core Tools** | File read / write / edit, glob search, grep search, shell execution |
| ๐ **Plugin Runtime** | Manifest-based plugins with hooks, aliases, virtual tools, and tool blocking |
| ๐ช **Nested Delegation** | Delegate subtasks to child agents with dependency-aware topological batching |
| ๐งฉ **Custom Agents** | Load local agent profiles from `~/.claude/agents` and `./.claude/agents`, inspect them via `/agents`, and delegate with `subagent_type` |
| ๐ก **Streaming** | Token-by-token streaming output with `--stream` |
| ๐ฌ **Slash Commands** | Local commands for context, config, account, search, MCP, remote, tasks, plan, hooks, and model control |
| ๐ **Remote Runtime** | Manifest-backed remote profiles with local `remote-mode`, `ssh-mode`, `teleport-mode`, and connect/disconnect state |
| ๐งญ **Task & Plan Runtime** | Persistent tasks and plans with sync, next-task selection, and blocked/unblocked state |
| ๐ฐ๏ธ **MCP Runtime** | Local MCP manifests plus real stdio MCP transport for resources and tools |
| ๐ **Search Runtime** | Provider-backed `web_search` plus provider activation and status reporting |
| โ๏ธ **Config & Account Runtime** | Local config mutation, settings inspection, account profiles, and login/logout state |
| ๐ **Ask-User Runtime** | Queued answer or interactive user-question flow with history tracking |
| ๐ฅ **Team Runtime** | Persisted local teams plus message history, handoff notes, and collaboration metadata |
| ๐ **Notebook Editing** | Native Jupyter notebook cell editing through `notebook_edit` |
| ๐ชต **Worktree Runtime** | Managed git worktrees with `worktree_enter`, `worktree_exit`, and live cwd switching |
| ๐งญ **Workflow Runtime** | Manifest-backed workflows with slash commands, CLI inspection, and recorded runs |
| โฐ **Remote Triggers** | Local remote triggers with create/update/run flows and npm-style trigger actions |
| ๐ช **Hook & Policy Runtime** | Trust reporting, safe env, managed settings, tool blocking, and budget overrides |
| ๐ง **LSP Code Intelligence** | Local LSP-style definitions, references, hover, symbols, diagnostics, and call hierarchy |
| ๐ง **Context Engine** | Automatic context building with CLAUDE.md discovery, compaction, and snipping |
| ๐ข **Tokenizer-Aware Accounting** | Model-aware token counting with cached tokenizer backends and fallback heuristics |
| ๐ **Prompt Budgeting** | Soft/hard prompt-window checks, token-budget reports, and preflight context collapse |
| ๐ **Session Persistence** | Save and resume agent sessions with file-history replay |
| ๐๏ธ **Background Sessions** | `agent-bg` and local daemon wrappers for background runs, logs, attach, and kill |
| ๐ฐ **Cost & Budget Control** | Token budgets, cost limits, tool-call caps, model-call caps |
| ๐ **Structured Output** | JSON schema response mode for programmatic use |
| ๐ **Permission System** | Granular control: `--allow-write`, `--allow-shell`, `--unsafe` |
| ๐๏ธ **OpenAI-Compatible** | Works with vLLM, Ollama, LiteLLM Proxy, OpenRouter โ any OpenAI-compatible API |
| ๐ **Qwen3-Coder** | First-class support for `Qwen3-Coder-30B-A3B-Instruct` via vLLM |
| ๐ฆ **Zero Dependencies** | Pure Python standard library โ nothing to install |
---
## ๐ Roadmap
### ๐ Documentation
| Document | Description |
|----------|-------------|
| [TESTING_GUIDE.md](TESTING_GUIDE.md) | Step-by-step commands to verify every feature |
| [PARITY_CHECKLIST.md](PARITY_CHECKLIST.md) | Full implementation status vs the npm source |
### โ
Done
- [x] Python CLI agent loop
- [x] Interactive chat mode (`agent-chat`) with multi-turn REPL
- [x] OpenAI-compatible local model backend
- [x] Qwen3-Coder support through vLLM with `qwen3_xml` tool parser
- [x] Ollama, LiteLLM Proxy, and OpenRouter backends
- [x] Core tools: `list_dir`, `read_file`, `write_file`, `edit_file`, `glob_search`, `grep_search`, `bash`
- [x] Context building and `/context`-style usage reporting
- [x] Slash commands: `/help`, `/context`, `/context-raw`, `/token-budget`, `/prompt`, `/permissions`, `/model`, `/tools`, `/agents`, `/memory`, `/status`, `/clear`
- [x] Session persistence and `agent-resume` flow
- [x] Permission system (read-only, write, shell, unsafe tiers)
- [x] Streaming token-by-token assistant output
- [x] Truncated-response continuation flow
- [x] Auto-snip and auto-compact context reduction
- [x] Reactive compaction retry on prompt-too-long errors
- [x] Preflight prompt-length validation and token-budget reporting
- [x] Preflight auto-compact/context collapse before backend prompt-too-long failures
- [x] Cost tracking and usage budget enforcement
- [x] Token, tool-call, model-call, and session-turn budgets
- [x] Structured output / JSON schema response mode
- [x] File history journaling with snapshot IDs and replay summaries
- [x] Nested agent delegation with dependency-aware topological batching
- [x] Agent manager with lineage tracking and group membership
- [x] Filesystem-backed custom agent profiles with built-in/user/project precedence
- [x] Local custom-agent create/update/delete flows via CLI and `/agents`
- [x] Local daemon-style background command family
- [x] Local background session workflows: `agent-bg`, `agent-ps`, `agent-logs`, `agent-attach`, `agent-kill`
- [x] Local remote runtime: manifest discovery, profile listing, connect/disconnect persistence, and CLI/slash flows
- [x] Local hook and policy runtime with trust reporting, safe env, tool blocking, and budget overrides
- [x] Local config runtime: config discovery, effective settings, source inspection, and config mutation
- [x] Local LSP runtime: definitions, references, hover, symbols, diagnostics, and call hierarchy
- [x] Local account runtime: profile discovery, login/logout state, and account CLI/slash flows
- [x] Local ask-user runtime: queued answers, history, and ask-user CLI/slash flows
- [x] Local team runtime: persisted teams, team messages, and team CLI/slash flows
- [x] Local search runtime with provider discovery, activation, and provider-backed `web_search`
- [x] Local MCP runtime: manifest resources, stdio transport, MCP resources, and MCP tool calls
- [x] Local task and plan runtimes with plan sync and dependency-aware task execution
- [x] Notebook edit tool in the real Python tool registry
- [x] Local workflow runtime with workflow list/get/run tools and CLI/slash flows
- [x] Local remote trigger runtime with create/update/run flows and CLI/slash inspection
- [x] Local managed git worktree runtime with live cwd switching and worktree CLI/slash flows
- [x] Local web GUI (FastAPI + vanilla JS SPA) with chat, sessions browser, slash command palette, and live settings (`python -m src.gui`)
- [x] Tokenizer-aware context accounting with cached tokenizer backends and heuristic fallback
- [x] Plugin runtime: manifest discovery, hooks, aliases, virtual tools, tool blocking
- [x] Plugin lifecycle hooks: resume, persist, delegate phases
- [x] Plugin session-state persistence and resume restoration
- [x] Query engine facade driving the real Python runtime
- [x] Compaction metadata with lineage IDs and revision summaries
- [x] Extended runtime tools: `web_fetch`, `web_search`, `tool_search`, `sleep`
- [x] Unit tests for the Python runtime
- [x] `pyproject.toml` packaging with `setuptools`
### ๐ฒ In Progress
- [ ] Full MCP parity beyond the current stdio transport and local manifest/resource/tool support
- [ ] Full slash-command parity with npm runtime
- [ ] Full interactive REPL / TUI behavior
- [ ] Full tokenizer/chat-message framing parity beyond the current tokenizer-aware accounting
- [ ] Hooks system parity
- [ ] Real remote transport/runtime parity beyond the current local remote-profile runtime
- [ ] Voice and VIM modes
- [ ] Editor and platform integrations
- [ ] Background and team features
---
## ๐๏ธ Architecture
```text
claw-code/
โโโ README.md
โโโ TESTING_GUIDE.md # How to test every feature
โโโ PARITY_CHECKLIST.md # Implementation status vs npm source
โโโ pyproject.toml
โโโ .gitignore
โโโ images/
โ โโโ logo.png
โโโ src/ # Python implementation
โ โโโ main.py # CLI entry point & argument parsing
โ โโโ agent_runtime.py # Core agent loop (LocalCodingAgent)
โ โโโ agent_tools.py # Tool definitions & execution engine
โ โโโ agent_prompting.py # System prompt assembly
โ โโโ agent_registry.py # Built-in + filesystem-backed custom agent discovery
โ โโโ agent_context.py # Context building & CLAUDE.md discovery
โ โโโ agent_context_usage.py # Context usage estimation & reporting
โ โโโ agent_session.py # Session state management
โ โโโ agent_slash_commands.py # Local slash command processing
โ โโโ agent_manager.py # Nested agent lineage & group tracking
โ โโโ agent_types.py # Shared dataclasses & type definitions
โ โโโ openai_compat.py # OpenAI-compatible API client (streaming)
โ โโโ plugin_runtime.py # Plugin manifest, hooks, aliases, virtual tools
โ โโโ agent_plugin_cache.py # Plugin discovery & prompt injection cache
โ โโโ session_store.py # Session serialization & persistence
โ โโโ transcript.py # Transcript block export & mutation tracking
โ โโโ query_engine.py # Query engine facade & runtime orchestration
โ โโโ mcp_runtime.py # Local MCP discovery and stdio MCP transport
โ โโโ search_runtime.py # Search providers and provider-backed web_search
โ โโโ remote_runtime.py # Local remote profiles, connect/disconnect state, remote CLI support
โ โโโ background_runtime.py # Local background sessions and daemon support
โ โโโ account_runtime.py # Local account profiles, login/logout state, account CLI support
โ โโโ ask_user_runtime.py # Local ask-user queued answers and interaction history
โ โโโ config_runtime.py # Local workspace config/settings discovery and mutation
โ โโโ lsp_runtime.py # Local LSP-style code intelligence and diagnostics
โ โโโ token_budget.py # Prompt-window budgeting and preflight prompt-length validation
โ โโโ plan_runtime.py # Persistent plan runtime and plan sync
โ โโโ task_runtime.py # Persistent task runtime and task execution
โ โโโ task.py # Task state model and task dataclasses
โ โโโ team_runtime.py # Local teams, messages, and collaboration metadata
โ โโโ workflow_runtime.py # Local workflow manifests and recorded workflow runs
โ โโโ remote_trigger_runtime.py # Local remote trigger manifests and trigger run history
โ โโโ worktree_runtime.py # Managed git worktree sessions and cwd switching
โ โโโ hook_policy.py # Hook/policy manifests, trust, and safe env handling
โ โโโ tokenizer_runtime.py # Tokenizer-aware context accounting backends
โ โโโ permissions.py # Tool permission filtering
โ โโโ cost_tracker.py # Cost & budget enforcement
โ โโโ commands.py # Mirrored command inventory
โ โโโ tools.py # Mirrored tool inventory
โ โโโ runtime.py # Mirrored runtime facade
โ โโโ reference_data/ # Mirrored inventory snapshots
โ โโโ gui/ # Local web GUI (FastAPI + vanilla JS SPA)
โ โโโ __main__.py # `python -m src.gui` entry point
โ โโโ server.py # FastAPI app and JSON endpoints
โ โโโ static/ # index.html, app.css, app.js
โโโ tests/ # Unit tests
โโโ test_agent_runtime.py
โโโ test_agent_context.py
โโโ test_agent_context_usage.py
โโโ test_agent_prompting.py
โโโ test_agent_slash_commands.py
โโโ test_main.py
โโโ test_query_engine_runtime.py
โโโ test_porting_workspace.py
```
---
## ๐ฆ Requirements
| Requirement | Details |
|-------------|---------|
| ๐ Python | `3.10` or higher |
| ๐ Dependencies | **None** โ pure Python standard library |
| ๐ฅ๏ธ Model Server | `vLLM`, `Ollama`, `LiteLLM Proxy`, or `OpenRouter`, with tool calling support |
| ๐ง Model | [`Qwen/Qwen3-Coder-30B-A3B-Instruct`](https://huggingface.co/Qwen/Qwen3-Coder-30B-A3B-Instruct) (recommended) |
---
## ๐ Quick Start
### 1. Start vLLM with Qwen3-Coder
vLLM must be started with automatic tool choice enabled. Use the `qwen3_xml` parser for Qwen3-Coder tool calling:
```bash
python -m vllm.entrypoints.openai.api_server \
--model Qwen/Qwen3-Coder-30B-A3B-Instruct \
--host 127.0.0.1 \
--port 8000 \
--enable-auto-tool-choice \
--tool-call-parser qwen3_xml
```
Verify the server is running:
```bash
curl http://127.0.0.1:8000/v1/models
```
> ๐ **References:** [vLLM Tool Calling Docs](https://docs.vllm.ai/en/v0.13.0/features/tool_calling/) ยท [OpenAI-Compatible Server](https://docs.vllm.ai/en/v0.13.0/serving/openai_compatible_server.html)
### Optional: Use Ollama Instead of vLLM
`claw-code-agent` can also work with Ollama because the runtime targets an OpenAI-compatible API. Use a model that supports tool calling well.
Example:
```bash
ollama serve
ollama pull qwen3
```
Then configure:
```bash
export OPENAI_BASE_URL=http://127.0.0.1:11434/v1
export OPENAI_API_KEY=ollama
export OPENAI_MODEL=qwen3
```
Notes:
- prefer tool-capable models such as `qwen3`
- plain chat-only models are not enough for full agent behavior
- Ollama does not use the `vLLM` parser flags shown above
> ๐ **References:** [Ollama OpenAI Compatibility](https://docs.ollama.com/api/openai-compatibility) ยท [Ollama Tool Calling](https://docs.ollama.com/capabilities/tool-calling)
### Optional: Use LiteLLM Proxy
`claw-code-agent` can also work through LiteLLM Proxy because the runtime targets an OpenAI-compatible chat completions API. The routed model still needs to support tool calling for full agent behavior.
Quick start example:
```bash
pip install 'litellm[proxy]'
litellm --model ollama/qwen3
```
LiteLLM Proxy runs on port `4000` by default. Then configure:
```bash
export OPENAI_BASE_URL=http://127.0.0.1:4000
export OPENAI_API_KEY=anything
export OPENAI_MODEL=ollama/qwen3
```
Notes:
- LiteLLM Proxy gives you an OpenAI-style gateway in front of many providers
- tool use still depends on the underlying routed model and provider behavior
- if you configure a LiteLLM master key, use that instead of `anything`
> ๐ **References:** [LiteLLM Docs](https://docs.litellm.ai/) ยท [LiteLLM Proxy Quick Start](https://docs.litellm.ai/)
### Optional: Use OpenRouter
`claw-code-agent` can also work with [OpenRouter](https://openrouter.ai/), a cloud API gateway that provides access to models from OpenAI, Anthropic, Google, Meta, and others through a single OpenAI-compatible endpoint. No local model server required.
Configure:
```bash
export OPENAI_BASE_URL=https://openrouter.ai/api/v1
export OPENAI_API_KEY=sk-or-v1-your-key-here
export OPENAI_MODEL=openai/gpt-4o-mini
```
Notes:
- sign up at [openrouter.ai](https://openrouter.ai/) and create an API key under [Keys](https://openrouter.ai/keys)
- model names use the `provider/model` format (e.g. `anthropic/claude-sonnet-4`, `openai/gpt-4o`, `google/gemini-2.5-pro`)
- tool calling support varies by model โ check the [model list](https://openrouter.ai/models) for capabilities
- this sends your conversation (including file contents and shell output) to OpenRouter and the upstream provider โ do not use with repos containing secrets or sensitive data
> ๐ **References:** [OpenRouter Docs](https://openrouter.ai/docs) ยท [Supported Models](https://openrouter.ai/models) ยท [API Keys](https://openrouter.ai/keys)
### 2. Configure Environment
```bash
export OPENAI_BASE_URL=http://127.0.0.1:8000/v1
export OPENAI_API_KEY=local-token
export OPENAI_MODEL=Qwen/Qwen3-Coder-30B-A3B-Instruct
```
### Use Another Model With vLLM
If you want to try another model, keep the same `vLLM` server setup and change the `--model` value when you launch `vLLM`.
Example:
```bash
python -m vllm.entrypoints.openai.api_server \
--model your-model-name \
--host 127.0.0.1 \
--port 8000 \
--enable-auto-tool-choice \
--tool-call-parser your_parser
```
Then update:
```bash
export OPENAI_MODEL=your-model-name
```
Notes:
- the documented path in this repository is `vLLM`
- the model must support tool calling well enough for agent use
- some model families require a different `--tool-call-parser`
- slash commands such as `/help`, `/context`, and `/tools` are local and do not require the model server
### 3. Run the Agent
```bash
# Read-only question
python3 -m src.main agent \
"Read src/agent_runtime.py and summarize how the loop works." \
--cwd .
# Write-enabled task
python3 -m src.main agent \
"Create TEST_QWEN_AGENT.md with one line: test ok" \
--cwd . --allow-write
# Shell-enabled task
python3 -m src.main agent \
"Run pwd and ls src, then summarize the result." \
--cwd . --allow-shell
# Interactive chat mode
python3 -m src.main agent-chat --cwd .
# Streaming output
python3 -m src.main agent \
"Explain the current architecture." \
--cwd . --stream
```
---
## ๐ ๏ธ Usage
### Agent Commands
| Command | Description |
|---------|-------------|
| `agent ` | Run the agent with a prompt |
| `agent-chat [prompt]` | Start interactive multi-turn chat mode |
| `agent-bg ` | Run the agent in a local background session |
| `agent-ps` | List local background sessions |
| `agent-logs ` | Show background session logs |
| `agent-attach ` | Show the current background output snapshot |
| `agent-kill ` | Stop a background session |
| `daemon ` | Daemon-style wrapper over local background sessions |
| `agent-prompt` | Show the assembled system prompt |
| `agent-context` | Show estimated context usage |
| `agent-context-raw` | Show the raw context snapshot |
| `token-budget` | Show prompt-window budget, reserves, and soft/hard input limits |
| `agents [agent_type]` | List active local agent definitions or show one agent profile |
| `agents-create ` | Create a project or user agent definition markdown file |
| `agents-update ` | Update an existing project or user agent definition |
| `agents-delete ` | Delete an existing project or user agent definition |
| `agent-resume ` | Resume a saved session |
### Runtime Utility Commands
| Command | Description |
|---------|-------------|
| `search-status` / `search-providers` / `search-activate` / `search` | Inspect and use the local search runtime |
| `mcp-status` / `mcp-resources` / `mcp-resource` / `mcp-tools` / `mcp-call-tool` | Inspect and use the local MCP runtime |
| `remote-status` / `remote-profiles` / `remote-disconnect` | Inspect local remote runtime state |
| `remote-mode` / `ssh-mode` / `teleport-mode` / `direct-connect-mode` / `deep-link-mode` | Activate local remote runtime modes |
| `config-status` / `config-effective` / `config-source` / `config-get` / `config-set` | Inspect and mutate local config/settings |
| `account-status` / `account-profiles` / `account-login` / `account-logout` | Inspect and mutate local account state |
### CLI Flags
| Flag | Description |
|------|-------------|
| `--cwd ` | Set the workspace directory |
| `--model ` | Override the model name |
| `--base-url ` | Override the API base URL |
| `--allow-write` | Allow the agent to modify files |
| `--allow-shell` | Allow the agent to execute shell commands |
| `--unsafe` | Allow destructive shell operations |
| `--stream` | Enable token-by-token streaming output |
| `--show-transcript` | Print the full message transcript |
| `--scratchpad-root ` | Override the scratchpad directory |
| `--system-prompt ` | Set a custom system prompt |
| `--append-system-prompt ` | Append to the system prompt |
| `--override-system-prompt ` | Replace the generated system prompt |
| `--add-dir ` | Add extra directories to context |
### Budget & Limit Flags
| Flag | Description |
|------|-------------|
| `--max-total-tokens ` | Total token budget |
| `--max-input-tokens ` | Input token budget |
| `--max-output-tokens ` | Output token budget |
| `--max-reasoning-tokens ` | Reasoning token budget |
| `--max-budget-usd ` | Maximum cost in USD |
| `--max-tool-calls ` | Maximum tool calls per run |
| `--max-delegated-tasks ` | Maximum delegated subtasks |
| `--max-model-calls ` | Maximum model API calls |
| `--max-session-turns ` | Maximum session turns |
| `--input-cost-per-million ` | Input token pricing |
| `--output-cost-per-million ` | Output token pricing |
### Context Control Flags
| Flag | Description |
|------|-------------|
| `--auto-snip-threshold ` | Auto-snip older messages at this token count |
| `--auto-compact-threshold ` | Auto-compact at this token count |
| `--compact-preserve-messages ` | Messages to preserve during compaction |
| `--disable-claude-md` | Disable CLAUDE.md discovery |
### Structured Output Flags
| Flag | Description |
|------|-------------|
| `--response-schema-file ` | JSON schema file for structured output |
| `--response-schema-name ` | Schema name identifier |
| `--response-schema-strict` | Enforce strict schema validation |
### Slash Commands
These are handled **locally** before the model loop:
| Command | Aliases | Description |
|---------|---------|-------------|
| `/help` | `/commands` | Show built-in slash commands |
| `/context` | `/usage` | Show estimated session context usage |
| `/context-raw` | `/env` | Show raw environment & context snapshot |
| `/token-budget` | `/budget` | Show prompt-window budget, reserves, and soft/hard input limits |
| `/mcp` | โ | Show MCP runtime status, tools, or a single MCP tool |
| `/resources` | โ | List MCP resources |
| `/resource` | โ | Read an MCP resource by URI |
| `/search` | โ | Show search status, providers, activate a provider, or run a search |
| `/remote` | โ | Show local remote status or activate a target |
| `/remotes` | โ | List local remote profiles |
| `/ssh` | โ | Activate an SSH-style remote profile |
| `/teleport` | โ | Activate a teleport-style remote profile |
| `/direct-connect` | โ | Activate a direct-connect remote profile |
| `/deep-link` | โ | Activate a deep-link remote profile |
| `/disconnect` | `/remote-disconnect` | Disconnect the active remote runtime target |
| `/account` | โ | Show account runtime status or profiles |
| `/login` | โ | Activate a local account profile or identity |
| `/logout` | โ | Clear the active account session |
| `/config` | `/settings` | Inspect effective config, sources, or a single config value |
| `/plan` | `/planner` | Show the local plan runtime state |
| `/tasks` | `/todo` | Show the local task list |
| `/task` | โ | Show a task by id |
| `/task-next` | `/next-task` | Show the next actionable tasks |
| `/prompt` | `/system-prompt` | Render the effective system prompt |
| `/hooks` | `/policy` | Show local hook/policy manifests |
| `/trust` | โ | Show trust mode, managed settings, and safe env values |
| `/permissions` | โ | Show active tool permission mode |
| `/model` | โ | Show or update the active model |
| `/tools` | โ | List registered tools with permission status |
| `/agents` | โ | List, show, create, update, or delete local agent definitions |
| `/memory` | โ | Show loaded CLAUDE.md memory bundle |
| `/status` | `/session` | Show runtime/session status summary |
| `/clear` | โ | Clear ephemeral runtime state |
```bash
python3 -m src.main agent "/help"
python3 -m src.main agent "/context" --cwd .
python3 -m src.main agent "/token-budget" --cwd .
python3 -m src.main agent "/tools" --cwd .
python3 -m src.main agent "/agents" --cwd .
python3 -m src.main agent "/status" --cwd .
```
### Custom Agent Definitions
Custom agent profiles can live in either of these directories:
- `./.claude/agents/*.md`
- `~/.claude/agents/*.md`
Project agents override user agents, and user agents override built-ins when the `agent_type` matches.
Example agent file:
```md
---
name: reviewer
description: "Review implementation changes carefully."
tools: read_file, grep_search
model: Qwen/Qwen3-Coder-30B-A3B-Instruct
initialPrompt: Start by identifying the highest-risk files.
---
Inspect code changes and summarize correctness risks, regressions, and missing tests.
```
Inspect the loaded profiles:
```bash
python3 -m src.main agents --cwd .
python3 -m src.main agents reviewer --cwd .
python3 -m src.main agent "/agents" --cwd .
python3 -m src.main agent "/agents show reviewer" --cwd .
```
Create, update, or delete agent files from the CLI:
```bash
python3 -m src.main agents-create reviewer \
--cwd . \
--description "Review implementation changes carefully." \
--prompt "Inspect code changes and summarize risks." \
--tools read_file,grep_search \
--model Qwen/Qwen3-Coder-30B-A3B-Instruct
python3 -m src.main agents-update reviewer \
--cwd . \
--description "Review implementation changes and tests carefully." \
--prompt "Focus on regressions, missing tests, and risky diffs."
python3 -m src.main agents-delete reviewer --cwd . --source project
```
Or use the local slash command management forms:
```bash
python3 -m src.main agent "/agents create reviewer :: Review implementation changes carefully. :: Inspect code changes and summarize risks." --cwd .
python3 -m src.main agent "/agents update reviewer Updated review description :: Focus on regressions and missing tests." --cwd .
python3 -m src.main agent "/agents delete reviewer" --cwd .
```
### Utility Commands
```bash
python3 -m src.main summary # Workspace summary
python3 -m src.main manifest # Workspace manifest
python3 -m src.main commands --limit 10 # Command inventory
python3 -m src.main tools --limit 10 # Tool inventory
```
---
## ๐ง Built-in Tools
The runtime currently includes core and extended tools:
| Tool | Description | Permission |
|------|-------------|------------|
| `list_dir` | List files and directories | ๐ข Always |
| `read_file` | Read file contents (with line ranges) | ๐ข Always |
| `write_file` | Write or create files | ๐ก `--allow-write` |
| `edit_file` | Edit files via exact string matching | ๐ก `--allow-write` |
| `glob_search` | Find files by glob pattern | ๐ข Always |
| `grep_search` | Search file contents by regex | ๐ข Always |
| `bash` | Execute shell commands | ๐ด `--allow-shell` |
| `web_fetch` | Fetch local or remote text content by URL | ๐ข Always |
| `search_status` / `search_list_providers` / `search_activate_provider` / `web_search` | Search runtime status and provider-backed web search | ๐ข Always |
| `tool_search` | Search the current Python tool registry | ๐ข Always |
| `sleep` | Bounded local wait tool | ๐ข Always |
| `config_list` / `config_get` / `config_set` | Inspect and mutate local workspace config | `config_set` is ๐ก `--allow-write` |
| `account_status` / `account_list_profiles` / `account_login` / `account_logout` | Inspect and mutate local account state | ๐ข Always |
| `remote_status` / `remote_list_profiles` / `remote_connect` / `remote_disconnect` | Inspect and mutate local remote runtime state | ๐ข Always |
| `mcp_list_resources` / `mcp_read_resource` / `mcp_list_tools` / `mcp_call_tool` | Use local MCP resources and transport-backed MCP tools | ๐ข Always |
| `plan_get` / `update_plan` / `plan_clear` | Inspect and mutate the local plan runtime | `update_plan` is ๐ก `--allow-write` |
| `task_next` / `task_list` / `task_get` / `task_create` / `task_update` / `task_start` / `task_complete` / `task_block` / `task_cancel` / `todo_write` | Persistent local task and todo management | write-like task mutations are ๐ก `--allow-write` |
| `delegate_agent` | Delegate work to nested child agents | ๐ข Always |
---
## ๐ Plugin System
Claw Code Agent supports a **manifest-based plugin runtime**. Drop a `plugin.json` in a `plugins/` subdirectory:
```json
{
"name": "my-plugin",
"hooks": {
"beforePrompt": "Inject guidance into the system prompt.",
"afterTurn": "Run after each agent turn.",
"onResume": "Reapply state on session resume.",
"beforePersist": "Save state before session is saved.",
"beforeDelegate": "Inject guidance before child agents.",
"afterDelegate": "Process child agent results."
},
"toolAliases": [
{ "name": "my_read", "baseTool": "read_file", "description": "Custom read alias." }
],
"virtualTools": [
{ "name": "my_tool", "description": "A virtual tool.", "responseTemplate": "result: {input}" }
]
}
```
> See [TESTING_GUIDE.md](TESTING_GUIDE.md) **Section 19** for full plugin testing commands.
---
## ๐ช Nested Agent Delegation
The agent can delegate subtasks to child agents with full context carryover:
```bash
python3 -m src.main agent \
"Delegate a subtask to inspect src/agent_runtime.py and return a summary." \
--cwd . --show-transcript
```
Features:
- Sequential and parallel subtask execution
- Dependency-aware topological batching
- Child-session save and resume
- Agent manager lineage tracking
> See [TESTING_GUIDE.md](TESTING_GUIDE.md) **Section 20** for delegation testing commands.
---
## ๐ฅ๏ธ Local Web GUI
If the terminal isn't your thing, launch the bundled browser GUI:
```bash
python3 -m src.gui --cwd . --allow-write --allow-shell
```
Your default browser opens to `http://127.0.0.1:8765` with a modern dark-themed chat UI.
| Flag | Description |
|------|-------------|
| `--host ` | Bind address (default `127.0.0.1`) |
| `--port ` | Port to listen on (default `8765`) |
| `--cwd ` | Workspace directory the agent operates in |
| `--model ` | Override the model name |
| `--base-url ` | Override the OpenAI-compatible API base URL |
| `--api-key ` | API key for the model server |
| `--session-dir ` | Where saved sessions live |
| `--allow-write` | Allow file write/edit tools |
| `--allow-shell` | Allow shell execution |
| `--temperature ` | Sampling temperature (default `0.0`) |
| `--timeout-seconds ` | Per-turn model timeout in seconds (default `120`) |
| `--stream` | Enable streaming model responses |
| `--max-turns ` | Per-run turn limit (default `12`) |
| `--max-budget-usd ` | Abort the run if total cost exceeds this |
| `--max-total-tokens ` | Token budget across prompt + completion |
| `--max-input-tokens ` | Input-token cap per call |
| `--max-output-tokens ` | Output-token cap per call |
| `--max-reasoning-tokens ` | Reasoning-token cap per call |
| `--max-tool-calls ` | Hard cap on tool invocations per run |
| `--max-model-calls ` | Hard cap on model invocations per run |
| `--max-delegated-tasks ` | Cap on nested delegated agents |
| `--max-session-turns ` | Cap across resumed sessions |
| `--system-prompt ` | Replace the rendered system prompt body |
| `--append-system-prompt ` | Append text to the rendered system prompt |
| `--override-system-prompt ` | Skip the default system prompt entirely and use this |
| `--response-schema-file ` | Load a structured-output schema from a JSON file |
| `--response-schema-name ` | Name the schema (default `response`) |
| `--response-schema-strict` | Reject responses that don't match the schema |
| `--auto-snip-threshold ` | Token threshold above which old messages are auto-snipped |
| `--auto-compact-threshold ` | Token threshold above which the conversation is auto-compacted |
| `--compact-preserve-messages ` | Number of recent messages preserved during a compact (default `4`) |
| `--disable-claude-md` | Skip discovery of `CLAUDE.md` files |
| `--add-dir ` | Additional working directory the agent may operate in (repeatable) |
| `--no-browser` | Don't auto-open a browser tab |
Every budget flag above is also editable at runtime through the **Budgets & limits** disclosure in the settings panel โ leave a field blank to clear the limit, type a number to set it.
The GUI surfaces:
- multi-turn chat with tool-call cards (collapsible JSON args + results)
- saved sessions sidebar with one-click resume
- slash command and skill pickers (`/` and `โ
` buttons, or `Cmd/Ctrl+K`)
- live settings panel (model, base URL, working dir, permissions)
- usage / cost meta in the composer footer
- pasted-content collapsing โ see below
- runtime knobs: temperature, timeout, streaming toggle, max turns
- a **Tasks** tab in the topbar โ list / create / start / complete / cancel against `.port_sessions/task_runtime.json`
### Paste large content
Paste anything โฅ500 characters into the composer (a logfile, a stack trace, an entire file) and the GUI replaces it with a short reference like `[Pasted text #1 +42 lines]`, plus a chip above the textarea showing `๐ [Pasted text #1] ยท 42 lines ยท 1894 chars ยท โ`.
- The reference stays editable โ type around it, delete it, or duplicate it; whatever survives at send-time is what gets expanded.
- The full content is held in the browser only and shipped with the next `/api/chat` request as `pasted_contents`.
- The server re-splices the original text back in before the agent runs, so the model sees the full payload โ never the placeholder.
- The chip's `โ` button drops both the content stash and any inline ref so it can't accidentally come along.
- The stash clears after every successful send and when you click `+ New chat`.
> **Note:** The GUI uses [FastAPI](https://fastapi.tiangolo.com/) and [Uvicorn](https://www.uvicorn.org/) under the hood. These get installed automatically if you install the package via `pip install -e .`. The core Python agent runtime itself remains dependency-free.
---
## ๐ Session Persistence
Each `agent` run automatically saves a resumable session:
```text
session_id=4f2c8c6f9c0e4d7c9c7b1b2a3d4e5f67
session_path=.port_sessions/agent/4f2c8c6f...
```
Resume a previous session:
```bash
python3 -m src.main agent-resume \
4f2c8c6f9c0e4d7c9c7b1b2a3d4e5f67 \
"Continue the previous task and finish the missing parts."
```
Resume directly into interactive chat:
```bash
python3 -m src.main agent-chat \
--resume-session-id \
--cwd .
```
Inspect saved sessions:
```bash
ls -lt .port_sessions/agent
```
> **Note:** Run `agent-resume` from the same `claw-code/` directory where the session was created. A resumed session continues from the saved transcript, not from scratch.
---
## ๐งช Testing
Run the full test suite:
```bash
python3 -m unittest discover -s tests -v
```
Smoke tests:
```bash
python3 -m src.main agent "/help"
python3 -m src.main agent-context --cwd .
python3 -m src.main agent \
"Read src/agent_session.py and summarize the message flow." \
--cwd .
```
> ๐ **Full testing guide:** See [TESTING_GUIDE.md](TESTING_GUIDE.md) for step-by-step commands covering the full implemented runtime surface.
---
## ๐ Permission Model
Claw Code Agent uses a **tiered permission system** to keep the agent safe by default:
| Tier | Capability | Flag Required |
|------|-----------|---------------|
| **Read-only** | List, read, glob, grep | None (default) |
| **Write** | + file creation and editing | `--allow-write` |
| **Shell** | + shell command execution | `--allow-shell` |
| **Unsafe** | + destructive shell operations | `--unsafe` |
---
## ๐ Parity Status
The full implementation checklist tracking parity against the npm `src` lives in [PARITY_CHECKLIST.md](PARITY_CHECKLIST.md).
It covers: core runtime, CLI modes, prompt assembly, context/memory, slash commands, tools, permissions, plugins, MCP, REPL/TUI, remote features, editor integrations, and internal subsystems.
---
## โ ๏ธ Disclaimer
- This repository is a **Python reimplementation** inspired by the Claude Code npm architecture.
- It does **not** ship the original npm source.
- It is **not** affiliated with or endorsed by Anthropic.
---
Built with ๐ Python ยท Powered by ๐ HarnessLab Team.