mcp-server

An MCP (Model Context Protocol) server (implemented in Go) that connects an LLM/agent to the Firebolt Data Warehouse. It exposes MCP tools for browsing Firebolt documentation, establishing Firebolt connections, executing SQL queries, and searching docs. It runs as a stdio-based MCP server by default, with optional SSE transport, and can be deployed via Docker or a binary. Authentication is done using Firebolt service-account credentials (client ID/secret) or via a Firebolt Core URL.

Evaluated Mar 30, 2026 (21d ago)
Homepage ↗ Repo ↗ Ai Ml mcp firebolt ai-agents data-warehouse sql documentation-access go sse stdio
⚙ Agent Friendliness
58
/ 100
Can an agent use this?
🔒 Security
60
/ 100
Is it safe for agents?
⚡ Reliability
32
/ 100
Does it work consistently?

Score Breakdown

⚙ Agent Friendliness

MCP Quality
78
Documentation
70
Error Messages
0
Auth Simplicity
70
Rate Limits
10

🔒 Security

TLS Enforcement
75
Auth Strength
65
Scope Granularity
25
Dep. Hygiene
55
Secret Handling
75

Authentication is based on Firebolt service-account client ID/secret or Core URL. The README shows secrets passed via CLI args/env variables (not in code snippets), but it does not document how secrets are stored, redacted, or logged, nor does it describe fine-grained tool authorization (all-or-nothing capabilities at the MCP level are implied). The server supports running with --network host in Docker examples, which can increase exposure if not constrained by your environment network policies. No CSP/SSRF/query-safety controls are documented; since it can execute SQL, additional guardrails are recommended when used with autonomous agents.

⚡ Reliability

Uptime/SLA
0
Version Stability
55
Breaking Changes
50
Error Recovery
25
AF Security Reliability

Best When

You want to integrate an MCP-capable LLM client (Cursor/VS Code Copilot Chat/Claude Desktop/others) with Firebolt for doc-aware SQL querying and analysis, and you can supply Firebolt service-account credentials (or a Firebolt Core URL).

Avoid When

You need a traditional REST/GraphQL API surface rather than MCP, or you require fine-grained, per-query authorization/permissions enforced at the MCP layer (not documented here).

Use Cases

  • Enable LLM/agent assistants to query Firebolt data warehouses via MCP tools
  • Let agents browse/search Firebolt SQL/function/documentation from within an MCP workflow
  • Connect multiple Firebolt engines/workspaces for agent-driven analysis
  • Provide “docs proof” gating before allowing connection/querying to reduce missing-context failures

Not For

  • Interactive, low-latency chat use cases that cannot tolerate extra steps (e.g., docs-read proof and/or doc/resource-heavy MCP calls)
  • Environments where exposing database query capabilities to an autonomous agent is not acceptable without additional guardrails
  • Clients that cannot integrate MCP over stdio or SSE

Interface

REST API
No
GraphQL
No
gRPC
No
MCP Server
Yes
SDK
No
Webhooks
No

Authentication

Methods: Firebolt service account (client ID + client secret via env/CLI args) Firebolt Core connection via CORE URL
OAuth: No Scopes: No

README indicates authentication is done via Firebolt service-account credentials or a Firebolt Core URL. No explicit per-scope authorization model is described for MCP tools.

Pricing

Free tier: No
Requires CC: No

No pricing details are provided for the MCP server itself; Firebolt usage costs (if any) would depend on the Firebolt plan.

Agent Metadata

Pagination
none
Idempotent
False
Retry Guidance
Not documented

Known Gotchas

  • Cursor may not work because the server uses MCP resources heavily; set FIREBOLT_MCP_DISABLE_RESOURCES=true to mitigate
  • The default firebolt_connect tool requires an LLM to present a read proof of Firebolt docs first; this can increase latency and token usage; disable via --skip-docs-proof / FIREBOLT_MCP_SKIP_DOCS_PROOF

Alternatives

Full Evaluation Report

Comprehensive deep-dive: security analysis, reliability audit, agent experience review, cost modeling, competitive positioning, and improvement roadmap for mcp-server.

AI-powered analysis · PDF + markdown · Delivered within 30 minutes

$99

Package Brief

Quick verdict, integration guide, cost projections, gotchas with workarounds, and alternatives comparison.

Delivered within 10 minutes

$3

Score Monitoring

Get alerted when this package's AF, security, or reliability scores change significantly. Stay ahead of regressions.

Continuous monitoring

$3/mo

Scores are editorial opinions as of 2026-03-30.

8642
Packages Evaluated
17761
Need Evaluation
586
Need Re-evaluation
Community Powered