Subgraph Registry

Discover and filter 15,500+ The Graph subgraphs by domain, network, or protocol with reliability scores and query URLs for precise data access.

This package is functionally coherent with a subgraph registry, but there are several practical mismatches you should consider before installing or providing secrets: - The skill metadata declares no required binaries or env vars, but SKILL.md/README expect you to run 'npx subgraph-registry-mcp' (so Node/npx is required) and the Python crawler/server will read a GATEWAY_API_KEY from .env if you provide one. Treat the Graph API key as sensitive — don't paste it into unknown packages. - On first run the MCP server will download a pre-built registry.db from a raw GitHub URL. That file comes from a third-party GitHub repo (PaulieB14). Downloading binary data from a raw URL is a supply-chain risk; verify the upstream repository and its release artifacts before trusting them. - The package runs network requests (queries to The Graph gateway, GitHub download). If you need to run it, prefer: (1) inspecting src/index.js and the npm package owner on the npm registry, (2) verifying the GitHub repo and commit history, (3) running in an isolated environment/container, (4) or building the registry locally by running the Python pipeline (python registry.py) rather than relying on the prebuilt DB. - If you plan to use this skill inside an agent that has access to secrets, avoid providing your Graph API key to the skill unless you have verified the package source and are comfortable with the risk. What would increase confidence: a declared install spec that lists Node/npx, an explicit requires.env entry for GATEWAY_API_KEY (or an explicit statement that no credentials are required for read-only use), and verification that the npm package and GitHub repo are owned by a known/trusted maintainer.

Type: OpenClaw Skill Name: subgraph-registry Version: 0.3.0 The Subgraph Registry skill bundle is a legitimate tool designed to help AI agents discover and analyze over 15,500 subgraphs on The Graph Network. The codebase includes a robust Python-based classification engine (classifier.py), a crawler for fetching on-chain metadata (crawler.py), and an MCP server implementation in both Node.js (src/index.js) and Python (mcp_server.py). The server facilitates data access by downloading a pre-computed SQLite database from the project's official GitHub repository, which is a standard practice for handling large datasets in MCP tools. No evidence of data exfiltration, malicious execution, or prompt injection was found; the use of environment variables for API keys and network access is strictly aligned with the stated purpose of querying The Graph's decentralized infrastructure.