MCP server for agent frameworks
Bolvrk as a first-class tool inside Claude, Cursor, and agent frameworks: agents call the verifier directly and read structured findings. Self-hosted on every plan; hosted with the full corpus on Team and above.
A Model Context Protocol server that exposes the verifier as tools, so an agent inside Claude, Cursor or any MCP-aware framework calls it directly and reads the structured findings without shelling out. Every tool result carries the findings JSON contract, the same object bolvrk check --json prints, plus a text summary the model reads.
Two servers, one contract. npx bolvrk mcp runs on the machine the agent runs on, over stdio, with the rules the public CLI bundles: check (Postgres, SQLite beta, optional live schema from a database URL), secrets (text or files, always local) and explain. Every plan, Free included, nothing leaves the machine.
The hosted server at bolvrk.com/api/mcp authenticates with a team token and adds what the hosted service has: the full corpus, live schema through stored connections, the team policy, and the run log, every check an agent makes lands there with source "mcp". Its tools are check, policy, rules and rule. There is no secrets tool hosted, because a credential scan never leaves the machine. Team plan and above.
# self-hosted, every plan
claude mcp add bolvrk -- npx -y bolvrk mcp
# hosted, Team and above
claude mcp add --transport http bolvrk https://bolvrk.com/api/mcp --header "Authorization: Bearer blv_..."Skills
The public repository ships a skills library next to the server: short SKILL.md files that tell an agent when to reach for the verifier, check a migration before committing it, scan a diff or a log for secrets before pasting it anywhere, explain a backfill against a shadow database. Copy one into your agent's skills folder and the workflow comes with it.
The CLI is open source and needs no account: npx bolvrk check migration.sql