moonbridge 0.5.1__tar.gz → 0.6.0__tar.gz

moonbridge-0.6.0/.release-please-manifest.json

@@ -0,0 +1,3 @@
+{
+  ".": "0.6.0"
+}

{moonbridge-0.5.1 → moonbridge-0.6.0}/CHANGELOG.md

@@ -5,6 +5,20 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [0.6.0](https://github.com/misty-step/moonbridge/compare/moonbridge-v0.5.2...moonbridge-v0.6.0) (2026-02-03)
+
+
+### Features
+
+* add AgentResult dataclass for typed result handling ([#63](https://github.com/misty-step/moonbridge/issues/63)) ([75e369f](https://github.com/misty-step/moonbridge/commit/75e369f5c00dece7d770c2a4a0b0cb804c422f81))
+
+## [0.5.2](https://github.com/misty-step/moonbridge/compare/moonbridge-v0.5.1...moonbridge-v0.5.2) (2026-02-01)
+
+
+### Bug Fixes
+
+* handle ProcessLookupError on SIGKILL path in _terminate_process ([#58](https://github.com/misty-step/moonbridge/issues/58)) ([b52d745](https://github.com/misty-step/moonbridge/commit/b52d7452bdc1777d63c42a2ce70d6e290f9b0c54))
+
 ## [0.5.1](https://github.com/misty-step/moonbridge/compare/moonbridge-v0.5.0...moonbridge-v0.5.1) (2026-02-01)
 
 

{moonbridge-0.5.1 → moonbridge-0.6.0}/CLAUDE.md

@@ -4,7 +4,7 @@ This file provides guidance to Claude Code (claude.ai/code) when working with co
 
 ## What is Moonbridge
 
-An MCP server that spawns Kimi K2.5 agents from any MCP client. Enables parallel agent execution—run 10 approaches simultaneously for the cost of one Claude request.
+An MCP server that spawns AI coding agents from any MCP client. Supports multiple backends (Kimi, Codex) with parallel execution—run 10 approaches simultaneously for a fraction of the cost.
 
 ## Commands
 
@@ -79,7 +79,9 @@ The MCP library is also stubbed in conftest when not installed, enabling tests t
 | Variable | Default | Purpose |
 |----------|---------|---------|
 | `MOONBRIDGE_ADAPTER` | `kimi` | CLI backend to use (`kimi`, `codex`) |
-| `MOONBRIDGE_TIMEOUT` | `600` | Default timeout (30-3600s) |
+| `MOONBRIDGE_TIMEOUT` | `600` | Global timeout fallback (30-3600s) |
+| `MOONBRIDGE_CODEX_TIMEOUT` | `1800` | Codex-specific timeout (30min default) |
+| `MOONBRIDGE_KIMI_TIMEOUT` | `600` | Kimi-specific timeout (10min default) |
 | `MOONBRIDGE_MAX_AGENTS` | `10` | Max parallel agents |
 | `MOONBRIDGE_ALLOWED_DIRS` | (none) | Colon-separated directory allowlist |
 | `MOONBRIDGE_STRICT` | `false` | Exit if ALLOWED_DIRS unset |
@@ -88,6 +90,8 @@ The MCP library is also stubbed in conftest when not installed, enabling tests t
 | `MOONBRIDGE_KIMI_MODEL` | (none) | Kimi-specific model override |
 | `MOONBRIDGE_CODEX_MODEL` | (none) | Codex-specific model override |
 
+Timeout resolution order: tool param > adapter env var > adapter default > global env var > 600s.
+
 Model resolution order: tool param > adapter env var > global env var > CLI default.
 
 All model values are validated: whitespace is stripped, empty strings become None, and models starting with `-` are rejected (flag injection prevention).

{moonbridge-0.5.1 → moonbridge-0.6.0}/PKG-INFO

@@ -1,14 +1,14 @@
 Metadata-Version: 2.4
 Name: moonbridge
-Version: 0.5.1
-Summary: MCP server for spawning Kimi K2.5 agents
+Version: 0.6.0
+Summary: MCP server for spawning AI coding agents (Kimi, Codex, and more)
 Project-URL: Homepage, https://github.com/misty-step/moonbridge
 Project-URL: Repository, https://github.com/misty-step/moonbridge
 Project-URL: Issues, https://github.com/misty-step/moonbridge/issues
 Author-email: Phaedrus <hello@mistystep.io>
 License-Expression: MIT
 License-File: LICENSE
-Keywords: agent,ai,claude,kimi,mcp
+Keywords: agent,ai,claude,codex,kimi,mcp
 Classifier: Development Status :: 4 - Beta
 Classifier: Environment :: Console
 Classifier: Intended Audience :: Developers
@@ -31,7 +31,7 @@ Description-Content-Type: text/markdown
 
 **Your MCP client just got a team.**
 
-Spawn Kimi K2.5 agents from Claude Code, Cursor, or any MCP client. Run 10 approaches in parallel for the cost of one Claude request.
+Spawn AI coding agents from Claude Code, Cursor, or any MCP client. Run 10 approaches in parallel for a fraction of the cost.
 
 ```bash
 uvx moonbridge
@@ -39,10 +39,12 @@ uvx moonbridge
 
 ## Quick Start
 
-1. **Install Kimi CLI and authenticate:**
-   ```bash
-   uv tool install --python 3.13 kimi-cli && kimi login
-   ```
+1. **Install at least one supported CLI:**
+
+   | Adapter | Install | Authenticate |
+   |---------|---------|--------------|
+   | Kimi (default) | `uv tool install --python 3.13 kimi-cli` | `kimi login` |
+   | Codex | `npm install -g @openai/codex` | Set `OPENAI_API_KEY` |
 
 2. **Add to MCP config** (`~/.mcp.json`):
    ```json
@@ -94,7 +96,8 @@ export MOONBRIDGE_SKIP_UPDATE_CHECK=1
 |------|----------|
 | `spawn_agent` | Single task: "Write tests for auth.ts" |
 | `spawn_agents_parallel` | Go wide: 10 agents, 10 approaches, pick the best |
-| `check_status` | Verify Kimi CLI is installed and authenticated |
+| `check_status` | Verify the configured CLI is installed and authenticated |
+| `list_adapters` | Show available adapters and their status |
 
 ### Example: Parallel Exploration
 
@@ -117,7 +120,10 @@ Three approaches. One request. You choose the winner.
 | Parameter | Type | Required | Description |
 |-----------|------|----------|-------------|
 | `prompt` | string | Yes | Task description for the agent |
-| `thinking` | boolean | No | Enable reasoning mode (default: false) |
+| `adapter` | string | No | Backend to use: `kimi`, `codex` (default: `kimi`) |
+| `model` | string | No | Model override (e.g., `gpt-5.2-codex`) |
+| `thinking` | boolean | No | Enable reasoning mode (Kimi only) |
+| `reasoning_effort` | string | No | Reasoning budget: `low`, `medium`, `high`, `xhigh` (Codex only) |
 | `timeout_seconds` | integer | No | Override default timeout (30-3600) |
 
 **`spawn_agents_parallel`**
@@ -126,7 +132,10 @@ Three approaches. One request. You choose the winner.
 |-----------|------|----------|-------------|
 | `agents` | array | Yes | List of agent configs (max 10) |
 | `agents[].prompt` | string | Yes | Task for this agent |
-| `agents[].thinking` | boolean | No | Enable reasoning for this agent |
+| `agents[].adapter` | string | No | Backend for this agent |
+| `agents[].model` | string | No | Model override for this agent |
+| `agents[].thinking` | boolean | No | Enable reasoning (Kimi only) |
+| `agents[].reasoning_effort` | string | No | Reasoning budget (Codex only) |
 | `agents[].timeout_seconds` | integer | No | Timeout for this agent |
 
 ## Response Format
@@ -136,7 +145,7 @@ All tools return JSON with these fields:
 | Field | Type | Description |
 |-------|------|-------------|
 | `status` | string | `success`, `error`, `timeout`, `auth_error`, or `cancelled` |
-| `output` | string | stdout from Kimi agent |
+| `output` | string | stdout from the agent |
 | `stderr` | string\|null | stderr if any |
 | `returncode` | int | Process exit code (-1 for timeout/error) |
 | `duration_ms` | int | Execution time in milliseconds |
@@ -158,37 +167,60 @@ All tools return JSON with these fields:
 
 ## Troubleshooting
 
-### "Kimi CLI not found"
+### "CLI not found"
 
-Install the Kimi CLI:
+Install the CLI for your chosen adapter:
 
 ```bash
+# Kimi
 uv tool install --python 3.13 kimi-cli
 which kimi
+
+# Codex
+npm install -g @openai/codex
+which codex
 ```
 
 ### "auth_error" responses
 
-Authenticate with Kimi:
+Authenticate with your chosen CLI:
 
 ```bash
+# Kimi
 kimi login
+
+# Codex
+export OPENAI_API_KEY=sk-...
 ```
 
 ### Timeout errors
 
-Increase the timeout for long-running tasks:
+Adapters have sensible defaults: Codex=1800s (30min), Kimi=600s (10min).
+
+For exceptionally long tasks, override explicitly:
 
 ```json
-{"prompt": "...", "timeout_seconds": 1800}
+{"prompt": "...", "timeout_seconds": 3600}
 ```
 
-Or set a global default:
+Or set per-adapter defaults via environment:
 
 ```bash
-export MOONBRIDGE_TIMEOUT=1800
+export MOONBRIDGE_CODEX_TIMEOUT=2400  # 40 minutes
+export MOONBRIDGE_KIMI_TIMEOUT=900    # 15 minutes
 ```
 
+## Timeout Best Practices
+
+| Task Type | Recommended |
+|-----------|-------------|
+| Quick query, status | 60-180s |
+| Simple edits | 300-600s |
+| Feature implementation | 1200-1800s |
+| Large refactor | 1800-3600s |
+
+Priority resolution: explicit param > adapter env > adapter default > global env > 600s fallback
+
 ### "MOONBRIDGE_ALLOWED_DIRS is not set" warning
 
 By default, Moonbridge warns at startup if no directory restrictions are configured. This is expected for local development. For shared/production environments, set allowed directories:

{moonbridge-0.5.1 → moonbridge-0.6.0}/README.md

@@ -2,7 +2,7 @@
 
 **Your MCP client just got a team.**
 
-Spawn Kimi K2.5 agents from Claude Code, Cursor, or any MCP client. Run 10 approaches in parallel for the cost of one Claude request.
+Spawn AI coding agents from Claude Code, Cursor, or any MCP client. Run 10 approaches in parallel for a fraction of the cost.
 
 ```bash
 uvx moonbridge
@@ -10,10 +10,12 @@ uvx moonbridge
 
 ## Quick Start
 
-1. **Install Kimi CLI and authenticate:**
-   ```bash
-   uv tool install --python 3.13 kimi-cli && kimi login
-   ```
+1. **Install at least one supported CLI:**
+
+   | Adapter | Install | Authenticate |
+   |---------|---------|--------------|
+   | Kimi (default) | `uv tool install --python 3.13 kimi-cli` | `kimi login` |
+   | Codex | `npm install -g @openai/codex` | Set `OPENAI_API_KEY` |
 
 2. **Add to MCP config** (`~/.mcp.json`):
    ```json
@@ -65,7 +67,8 @@ export MOONBRIDGE_SKIP_UPDATE_CHECK=1
 |------|----------|
 | `spawn_agent` | Single task: "Write tests for auth.ts" |
 | `spawn_agents_parallel` | Go wide: 10 agents, 10 approaches, pick the best |
-| `check_status` | Verify Kimi CLI is installed and authenticated |
+| `check_status` | Verify the configured CLI is installed and authenticated |
+| `list_adapters` | Show available adapters and their status |
 
 ### Example: Parallel Exploration
 
@@ -88,7 +91,10 @@ Three approaches. One request. You choose the winner.
 | Parameter | Type | Required | Description |
 |-----------|------|----------|-------------|
 | `prompt` | string | Yes | Task description for the agent |
-| `thinking` | boolean | No | Enable reasoning mode (default: false) |
+| `adapter` | string | No | Backend to use: `kimi`, `codex` (default: `kimi`) |
+| `model` | string | No | Model override (e.g., `gpt-5.2-codex`) |
+| `thinking` | boolean | No | Enable reasoning mode (Kimi only) |
+| `reasoning_effort` | string | No | Reasoning budget: `low`, `medium`, `high`, `xhigh` (Codex only) |
 | `timeout_seconds` | integer | No | Override default timeout (30-3600) |
 
 **`spawn_agents_parallel`**
@@ -97,7 +103,10 @@ Three approaches. One request. You choose the winner.
 |-----------|------|----------|-------------|
 | `agents` | array | Yes | List of agent configs (max 10) |
 | `agents[].prompt` | string | Yes | Task for this agent |
-| `agents[].thinking` | boolean | No | Enable reasoning for this agent |
+| `agents[].adapter` | string | No | Backend for this agent |
+| `agents[].model` | string | No | Model override for this agent |
+| `agents[].thinking` | boolean | No | Enable reasoning (Kimi only) |
+| `agents[].reasoning_effort` | string | No | Reasoning budget (Codex only) |
 | `agents[].timeout_seconds` | integer | No | Timeout for this agent |
 
 ## Response Format
@@ -107,7 +116,7 @@ All tools return JSON with these fields:
 | Field | Type | Description |
 |-------|------|-------------|
 | `status` | string | `success`, `error`, `timeout`, `auth_error`, or `cancelled` |
-| `output` | string | stdout from Kimi agent |
+| `output` | string | stdout from the agent |
 | `stderr` | string\|null | stderr if any |
 | `returncode` | int | Process exit code (-1 for timeout/error) |
 | `duration_ms` | int | Execution time in milliseconds |
@@ -129,37 +138,60 @@ All tools return JSON with these fields:
 
 ## Troubleshooting
 
-### "Kimi CLI not found"
+### "CLI not found"
 
-Install the Kimi CLI:
+Install the CLI for your chosen adapter:
 
 ```bash
+# Kimi
 uv tool install --python 3.13 kimi-cli
 which kimi
+
+# Codex
+npm install -g @openai/codex
+which codex
 ```
 
 ### "auth_error" responses
 
-Authenticate with Kimi:
+Authenticate with your chosen CLI:
 
 ```bash
+# Kimi
 kimi login
+
+# Codex
+export OPENAI_API_KEY=sk-...
 ```
 
 ### Timeout errors
 
-Increase the timeout for long-running tasks:
+Adapters have sensible defaults: Codex=1800s (30min), Kimi=600s (10min).
+
+For exceptionally long tasks, override explicitly:
 
 ```json
-{"prompt": "...", "timeout_seconds": 1800}
+{"prompt": "...", "timeout_seconds": 3600}
 ```
 
-Or set a global default:
+Or set per-adapter defaults via environment:
 
 ```bash
-export MOONBRIDGE_TIMEOUT=1800
+export MOONBRIDGE_CODEX_TIMEOUT=2400  # 40 minutes
+export MOONBRIDGE_KIMI_TIMEOUT=900    # 15 minutes
 ```
 
+## Timeout Best Practices
+
+| Task Type | Recommended |
+|-----------|-------------|
+| Quick query, status | 60-180s |
+| Simple edits | 300-600s |
+| Feature implementation | 1200-1800s |
+| Large refactor | 1800-3600s |
+
+Priority resolution: explicit param > adapter env > adapter default > global env > 600s fallback
+
 ### "MOONBRIDGE_ALLOWED_DIRS is not set" warning
 
 By default, Moonbridge warns at startup if no directory restrictions are configured. This is expected for local development. For shared/production environments, set allowed directories:

{moonbridge-0.5.1 → moonbridge-0.6.0}/pyproject.toml

@@ -1,14 +1,14 @@
 [project]
 name = "moonbridge"
 dynamic = ["version"]
-description = "MCP server for spawning Kimi K2.5 agents"
+description = "MCP server for spawning AI coding agents (Kimi, Codex, and more)"
 readme = "README.md"
 requires-python = ">=3.11"
 license = "MIT"
 authors = [
     {name = "Phaedrus", email = "hello@mistystep.io"}
 ]
-keywords = ["mcp", "kimi", "claude", "agent", "ai"]
+keywords = ["mcp", "kimi", "codex", "claude", "agent", "ai"]
 classifiers = [
     "Development Status :: 4 - Beta",
     "Environment :: Console",

{moonbridge-0.5.1 → moonbridge-0.6.0}/src/moonbridge/__init__.py

@@ -1,8 +1,8 @@
-"""MCP server for spawning Kimi K2.5 agents."""
+"""MCP server for spawning AI coding agents."""
 
 from __future__ import annotations
 
-__version__ = "0.5.1"
+__version__ = "0.6.0"
 
 from .server import main, run, server
 

{moonbridge-0.5.1 → moonbridge-0.6.0}/src/moonbridge/adapters/base.py

@@ -1,5 +1,5 @@
 from dataclasses import dataclass
-from typing import Protocol
+from typing import Any, Protocol
 
 
 @dataclass(frozen=True)
@@ -18,6 +18,35 @@ class AdapterConfig:
     default_timeout: int = 600
 
 
+@dataclass(frozen=True)
+class AgentResult:
+    """Agent execution result."""
+
+    status: str
+    output: str
+    stderr: str | None
+    returncode: int
+    duration_ms: int
+    agent_index: int
+    message: str | None = None
+    raw: dict[str, Any] | None = None
+
+    def to_dict(self) -> dict[str, Any]:
+        payload: dict[str, Any] = {
+            "status": self.status,
+            "output": self.output,
+            "stderr": self.stderr,
+            "returncode": self.returncode,
+            "duration_ms": self.duration_ms,
+            "agent_index": self.agent_index,
+        }
+        if self.message is not None:
+            payload["message"] = self.message
+        if self.raw is not None:
+            payload["raw"] = self.raw
+        return payload
+
+
 class CLIAdapter(Protocol):
     """Protocol for CLI backend adapters."""
 

{moonbridge-0.5.1 → moonbridge-0.6.0}/src/moonbridge/adapters/codex.py

@@ -55,6 +55,7 @@ class CodexAdapter:
             "gpt-5.1-codex-mini",
             "gpt-5.1-codex-max",
         ),
+        default_timeout=1800,  # 30 minutes - Codex runs long
     )
 
     def build_command(

{moonbridge-0.5.1 → moonbridge-0.6.0}/src/moonbridge/adapters/kimi.py

@@ -38,6 +38,7 @@ class KimiAdapter:
         install_hint="uv tool install kimi-cli",
         supports_thinking=True,
         known_models=("kimi-k2.5",),
+        default_timeout=600,  # 10 minutes - Kimi is faster
     )
 
     def build_command(