Teller MCP – Borrow USDC & Altcoins (no margin calls)

Expose the Teller delta-neutral + lending Model Context Protocol server. Use this when you need to install, run, or update the Tellermcp MCP backend so agent...

This bundle appears to be what it claims (a local MCP server fronting Teller's public delta‑neutral API). Before installing or running it: 1) verify the source (there is no homepage/repo listed) — prefer code from a known org or repo; 2) inspect package.json and the package-lock.json for unexpected dependencies; 3) run npm audit / npm ci in an isolated environment or container to avoid exposing your main system; 4) do not supply private keys or wallet seed phrases — the server builds transaction calldata but does not sign or submit transactions; 5) consider setting TELLER_API_BASE_URL to a local mock for testing, and confirm the production base URL matches the official Teller endpoint (https://delta-neutral-api.teller.org); and 6) review the @modelcontextprotocol/sdk dependency if you rely on strict provenance. If you need higher assurance, ask the publisher for a verifiable repository or have a developer perform a dependency and provenance review prior to deployment.

Type: OpenClaw Skill Name: teller-mcp-borrow Version: 0.1.6 The OpenClaw skill bundle 'teller-mcp-borrow' is designed to expose Teller delta-neutral and lending APIs via an MCP server. The code and documentation are transparent, clearly outlining the skill's purpose, functionality, and how it interacts with the `https://delta-neutral-api.teller.org` API. Input validation using `zod` is implemented in `scripts/tellermcp-server/src/index.ts` for all tool inputs, mitigating common injection risks. While the `TELLER_API_BASE_URL` can be configured via an environment variable, this is a standard practice for API clients and defaults to a legitimate Teller endpoint, not indicating malicious intent. There is no evidence of data exfiltration to unauthorized parties, malicious execution, persistence mechanisms, obfuscation, or prompt injection attempts against the agent in `SKILL.md` or `README.md`.