← 返回 Skills 市场

Cerebrun

Name: Cerebrun
Author: niyoseris

作者 niyoseris · GitHub ↗ · v1.0.1

cross-platform ⚠ suspicious

494

总下载

当前安装

版本数

在 OpenClaw 中安装

/install cerebrun

功能描述

MCP client for Cerebrun - comprehensive personal context and memory management system. Retrieve user context layers (language, projects, identity, vault), pe...

使用说明 (SKILL.md)

Cerebrun MCP Client

Cerebrun (cereb.run) is a Model Context Protocol (MCP) server that provides persistent personal context management across sessions.

Quick Start

All requests require:

api_key: Cerebrun API key (Bearer token)
base_url: https://cereb.run/mcp

Context Layers

Layer 0 - Language, timezone, comms prefs Layer 1 - Projects, goals, pinned memories Layer 2 - Personal identity info, API and other keys Layer 3 - Encrypted vault (requires consent)

Usage

Get Context

curl -X POST
-H "Authorization: Bearer $API_KEY"
-H "Content-Type: application/json"
-d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"get_context","arguments":{"layer":0}}}'
https://cereb.run/mcp

Search Context

curl -X POST
-H "Authorization: Bearer $API_KEY"
-d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"search_context","arguments":{"query":"Rust authentication","limit":5}}}'
https://cereb.run/mcp

Push Knowledge

curl -X POST
-H "Authorization: Bearer $API_KEY"
-d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"push_knowledge","arguments":{"content":"Important insight","category":"learning","tags":["rust","performance"]}}}'
https://cereb.run/mcp

Chat via Gateway

curl -X POST
-H "Authorization: Bearer $API_KEY"
-d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"chat_with_llm","arguments":{"message":"Hello","provider":"openai","model":"gpt-4"}}}'
https://cereb.run/mcp

Tools Reference

See REFERENCES.md for complete API documentation.

Script Usage

scripts/cerebrun.py get_context --layer 0 --api-key YOUR_KEY
scripts/cerebrun.py search_context --query "project" --api-key YOUR_KEY
scripts/cerebrun.py push_knowledge --content "New idea" --category "todo" --api-key YOUR_KEY

Configuration

Store API key in environment: CEREBRUN_API_KEY or pass via --api-key

安全使用建议

This skill is a legitimate-looking client for the remote service cereb.run, but the package metadata fails to declare that it requires an API key. Before installing: (1) verify the publisher/source (homepage is missing), (2) expect to provide a Cerebrun bearer token — avoid reusing highly privileged tokens, (3) understand that using the skill can return highly sensitive data (identity fields, stored API keys in layer 2, and vault data via explicit requests), so only enable it if you trust the Cerebrun service and its privacy/security practices, (4) ask the publisher to correct the metadata to declare the required env var/primary credential, and (5) restrict the token you give to least privilege (scoped token) and monitor usage. If you cannot verify the service/publisher, treat this skill as risky and do not provide real secrets.

功能分析

Type: OpenClaw Skill Name: cerebrun Version: 1.0.1 The OpenClaw skill 'cerebrun' is a client for the Cerebrun Model Context Protocol (MCP) server, designed to manage personal context, knowledge, and LLM interactions. The `SKILL.md` and `scripts/cerebrun.py` clearly outline its purpose, which includes handling sensitive data like personal identity and an encrypted vault (Layer 2 and 3 context). The Python script uses standard `urllib.request` to communicate with the designated `https://cereb.run/mcp` endpoint, retrieves API keys securely from environment variables or CLI arguments, and does not exhibit any signs of unauthorized data exfiltration, malicious execution, persistence mechanisms, or prompt injection attempts against the agent. All actions are aligned with the stated purpose of managing personal data on the Cerebrun platform.

能力评估

⚠ Purpose & Capability

The skill's name/description match the included code and SKILL.md: it is an MCP client for cereb.run. However the registry metadata declares no required environment variables or primary credential even though the SKILL.md and scripts clearly require an API key (CEREBRUN_API_KEY / --api-key). That mismatch is a material incoherence: a networked client that accesses sensitive user layers should declare its credential requirements in metadata.

⚠ Instruction Scope

SKILL.md and scripts instruct the agent to fetch/update context layers (including layer 2: identity/API keys and layer 3: encrypted vault via explicit request), search knowledge, and fetch full conversation histories. Those operations are within the stated purpose, but they allow access to highly sensitive data and conversation history. The instructions also show how to send messages to external LLM providers via the Gateway. There are no unexpected local file or system-access instructions, but the scope includes potentially exfiltratable user secrets via the remote API.

✓ Install Mechanism

No install spec is present and the skill is instruction+script only. The included Python script is readable, uses only standard library urllib, and points to a single well-formed endpoint (https://cereb.run/mcp). There are no downloads, extract steps, or third-party package installs.

⚠ Credentials

The code and README expect an API key (CEREBRUN_API_KEY or --api-key), but the registry's required env/primaryEnv fields are empty. That omission is disproportionate and important: the skill needs a bearer token to operate and will use it to access context that may include other secrets. The skill does not request unrelated local secrets, but the ability to retrieve layer 2 (which the SKILL.md says may contain API/other keys) increases sensitivity of the single required credential.

✓ Persistence & Privilege

The skill is not always-enabled and does not request elevated platform privileges. It does not modify other skills or system-wide config. Autonomous invocation is allowed (default) — which is normal — but not combined here with always:true or other red flags.

如何使用

确保已安装 OpenClaw（本地或 Docker 部署）
在对话框中输入安装命令：/install cerebrun
安装完成后，直接呼叫该 Skill 的名称或使用 /cerebrun 触发
根据 Skill 的参数说明提供必要输入，即可获得结构化输出

版本历史

v1.0.1

- Updated Layer 2 description in SKILL.md to mention storage of "API and other keys" alongside personal identity info.

v1.0.0

Initial release of the Cerebrun MCP client. - Provides tools for managing personal context and memory via the Cerebrun MCP server. - Supports retrieving user preferences, searching personal knowledge base, managing projects/goals, and handling personal identity and encrypted vault info. - Includes semantic search and LLM Gateway chat integrations. - Offers Python script examples and curl usage for all core functions. - API key configuration instructions provided.

元数据

Slug cerebrun

版本 1.0.1

许可证 —

累计安装 0

当前安装数 0

历史版本数 2

常见问题

Cerebrun 是什么？

MCP client for Cerebrun - comprehensive personal context and memory management system. Retrieve user context layers (language, projects, identity, vault), pe... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 494 次。

如何安装 Cerebrun？

在 OpenClaw 或 Claude Code 对话框中运行命令「/install cerebrun」即可一键安装，无需额外配置。

Cerebrun 是免费的吗？

是的，Cerebrun 完全免费（开源免费），可自由下载、安装和使用。

Cerebrun 支持哪些平台？

Cerebrun 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。

谁开发了 Cerebrun？

由 niyoseris（@niyoseris）开发并维护，当前版本 v1.0.1。