← Back to Skills Marketplace
271
Downloads
0
Stars
0
Active Installs
3
Versions
Install in OpenClaw
/install predictfun-mcp
Description
Access Predict.fun prediction market data on BNB Chain — platform stats, market analysis, trader profiling, yield mechanics, and behavioral meta-tools via Th...
Usage Guidance
This package appears to do what it says: query three Predict.fun subgraphs through The Graph Gateway and optionally run a local MCP (SSE/HTTP) server. Before installing: (1) only provide a The Graph API key (GRAPH_API_KEY); use a limited or monitoring-enabled key if possible because queries are billed to it; (2) running with --http or --http-only opens a local port (default 3850) — avoid exposing that port to untrusted networks or firewall it; (3) npx will fetch the package from the npm registry — verify the package and maintainer on npm/GitHub (the repo URL is in package.json) and confirm version consistency if that matters; (4) if you need stricter assurance, review the published npm package contents and the dist/index.js source included here (no obfuscated or exfiltrating code was found).
Capability Analysis
Type: OpenClaw Skill
Name: predictfun-mcp
Version: 0.5.2
The predictfun-mcp skill bundle is a legitimate MCP server designed to provide structured access to Predict.fun prediction market data on the BNB Chain via The Graph. The implementation in src/index.ts and dist/index.js uses the standard Model Context Protocol SDK to define tools for querying subgraphs related to orderbooks, positions, and yield mechanics. All network activity is directed to the official Graph Gateway (gateway.thegraph.com), and the required GRAPH_API_KEY is used appropriately for these requests. No evidence of data exfiltration, malicious execution, or prompt injection was found.
Capability Assessment
Purpose & Capability
Name/description, required binary (node), and required env var (GRAPH_API_KEY) all align with a Node-based MCP server that queries The Graph Gateway for Predict.fun subgraphs. The hard-coded subgraph IDs and Graph gateway endpoints fit the described functionality.
Instruction Scope
SKILL.md and the shipped code instruct the agent to perform GraphQL queries against gateway.thegraph.com using the provided API key, start an optional local HTTP/SSE server, and return structured results. The instructions do not read arbitrary local files, request other environment variables, or send data to unexpected external endpoints.
Install Mechanism
No high-risk download URLs are used; the documented install/run method is 'npx predictfun-mcp' (standard npm usage). The package.json dependencies (@modelcontextprotocol/sdk, express) are consistent with an MCP server. There is no custom installer or obscure remote fetch in the code provided.
Credentials
Only GRAPH_API_KEY is required and is directly used to authenticate requests to The Graph Gateway. No unrelated secrets, system credentials, or config paths are requested.
Persistence & Privilege
The skill does not request 'always: true' or other elevated persistent privileges. It runs as a user process and optionally opens a local HTTP/SSE port (default 3850); this is expected for an MCP server but should be considered when exposing to networks.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install predictfun-mcp - After installation, invoke the skill by name or use
/predictfun-mcp - Provide required inputs per the skill's parameter spec and get structured output
Version History
v0.5.2
Add OpenClaw YAML frontmatter with GRAPH_API_KEY gating for agent discovery
v0.5.1
- Added requirements section detailing Node.js version and the need for a GRAPH_API_KEY from The Graph Studio.
- Installation instructions now include setting the GRAPH_API_KEY environment variable.
- New section explains network and data behavior, including subgraph usage and server port configuration.
- Removed package-lock.json from the repository.
v0.5.0
- Major update: Adds comprehensive data access and analysis tools for Predict.fun on BNB Chain.
- New endpoints for platform stats, market rankings, trader profiling, and behavioral meta-analysis.
- Includes yield tracking with Venus Protocol integration and custom GraphQL querying.
- Supports detailed market structure tagging and searching by archetype or structure.
- New use cases: real-time stats, yield mechanics monitoring, and advanced trader/market profiling.
Metadata
Frequently Asked Questions
What is Predict.fun MCP?
Access Predict.fun prediction market data on BNB Chain — platform stats, market analysis, trader profiling, yield mechanics, and behavioral meta-tools via Th... It is an AI Agent Skill for Claude Code / OpenClaw, with 271 downloads so far.
How do I install Predict.fun MCP?
Run "/install predictfun-mcp" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Predict.fun MCP free?
Yes, Predict.fun MCP is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Predict.fun MCP support?
Predict.fun MCP is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Predict.fun MCP?
It is built and maintained by PaulieB14 (@paulieb14); the current version is v0.5.2.
More Skills