ShibaClaw

data-ai MCP Server

🐕 Self-hosted security-first AI agent · 22 providers · 11 chat channels · WebUI · 3-level memory · cron · heartbeat · skills · MCP

Verified
data-aidata-ai
8 views90 stars7 forksApache-2.0

Why This Matters

Discovered via agent-topic:ai-agent and last synced 3mo ago.

Verified
Source
agent-topic:ai-agent
Stars
90
Last synced
3mo ago
Install
Check source

Install

Install instructions not detected yet

Check the source repository for the latest setup steps.

View source instructions
33
Tools
0
Resources
0
Prompts
Standard I/O
Transport

Available Tools (33)

web_fetch

HTTP fetch with SSRF protection, DNS rebinding defense, and redirect validation

DeepSeek

`DEEPSEEK_API_KEY`

shibaclaw-web

WebUI (Starlette + native WebSocket), cron service

Problem

Try

web_search

Brave, Tavily, SearXNG, Jina, or DuckDuckGo (fallback, no key needed)

Anthropic

`ANTHROPIC_API_KEY`

shibaclaw-gateway

Core agent loop, message bus, channel integrations

WebUI

~120 MB

Feature

ShibaClaw

exec

Shell commands with 20+ deny-pattern guards, encoding normalization, and CVE scanning

cron

Schedule one-time or recurring jobs (cron expressions, intervals, ISO dates, timezone-aware)

OpenAI

`OPENAI_API_KEY`

MiniMax

`MINIMAX_API_KEY`

Real-time

Native WebSocket (`/ws` on WebUI, port `19998` on gateway)

memory_search

Ranked search over session history (TF-IDF + recency + importance scoring)

Groq

`GROQ_API_KEY`

Layer

Technology

Server

Uvicorn → Starlette (ASGI)

OpenRouter

PKCE browser flow, stores returned API key in provider config

Component

Idle

ZeroClaw

Nanobot

spawn

Optional background worker for a focused task; reports back to the main session when done

DASHSCOPE_API_KEY

¹ Setting `GEMINI_API_KEY` in the environment is sufficient — no stored key required. The Google OpenAI-compatible endpoint is pre-configured. ### Gateway / Proxy OpenRouter · AiHubMix · SiliconFlow · VolcEngine · BytePlus — auto-detected by key prefix or `api_base`. ### Local Ollama (`http://localhost:11434`) · LM Studio · llama.cpp · vLLM · any OpenAI-compatible endpoint(`http://localhost:1234/v1`) > **Note for Docker users:** If you run ShibaClaw via Docker Compose, `localhost` points inside the container itself. To connect to a local server running on your host machine (like LM Studio or Ollama on Windows/Mac), use: > `http://host.docker.internal:1234/v1` (or `11434` for Ollama). On native Linux, use `http://172.17.0.1:port`. ### OAuth

Frontend

Vanilla JS · Marked.js · Highlight.js

MCP

Connect any MCP server (stdio, SSE, or streamable HTTP) — tools auto-registered as `mcp_<server>_<tool>`

Sessions

JSONL append-only per session (cache-friendly for LLM prompt prefixes)

Tool

What it does

message

Cross-channel messaging with media attachments

Moonshot

`MOONSHOT_API_KEY`

Flow

Setup

Provider

Env Variable

Service

Role

Gateway

~120 MB