mcpdirectory.dev
9,129 servers · updated 23 Sept

Solvscore MCP server

community Active last commit 28 d
{
  "mcpServers": {
    "solvscore": {
      "type": "http",
      "url": "https://solvscore.com/mcp"
    }
  }
}
file .mcp.json (in your project root)version 1.1.0

What it does

SolvScore provides credit scoring and underwriting for AI agents based on verified work history and reputation bonds rather than traditional assets. Developers use it to assess creditworthiness of unknown agents before extending credit, prepayment, or hiring them for work.

  • Check an agent's credit score and borrowing limit before deciding whether to extend credit or accept prepayment from them.
  • Underwrite a specific loan request to an agent and receive an approve/decline decision with terms, APR, and a reason.
  • Register your own agent and link its work history to build its credit file and reputation bonds over time.

Tools this server exposes

ToolDoes
get_agent_credit_scoreReturns trust score (0-100), bonds held, credit limit, APR, and red flags (sybil flagged, cooling period) for an agent.
underwrite_agent_loanReturns an approve or decline decision for a specific loan amount, including limit, APR, tier, and plain-language reason.
register_agentEnrolls an agent in the credit system and returns a credit file ID; optionally links an existing wallet with verified history.
get_agent_attestationsReturns third-party attestations about a wallet onchain, including Coinbase Verified Account identity proofs if present.
get_credit_market_statsReturns aggregate market data including total capacity, bonds outstanding and slashed, and quotes issued versus approved.
get_credit_leaderboardReturns the most creditworthy agents by score, useful for discovering potential counterparties.
get_scoring_weightsReturns the governance-set weights that determine how each agent's credit score is calculated.

Before you enable it

An MCP server runs with whatever access you give it. Start with a development account rather than production, and check which tools the server exposes before you let an agent call them unattended.