MCP Tools DirectoryAbout this integration
Pre-flight query cost and result-size guardrails for AI agents on BigQuery, Snowflake, Databricks
- Transport
- stdio
- Authentication
- unknown
- Initial setup
- credentials
- Runtime
- unattended
- Evidence
- documented
- Version
- 0.1.1
- Package
- cost-guard-mcp
- Last compatibility test
- Not independently tested
Connect your agent
cost-guard-mcpPublisher README and Python manifest reviewed from pinned captures. Warehouse accuracy and safety behavior are documented claims; no credential, database, or query was used.
Capabilities: Check warehouse credentials, Describe engine capabilities, Estimate query cost, Run a query with explicit bounds
Connected profiles
Additional details
io.github.mcpsmiths/cost-guard-mcp
Source ↗ · Checked 2026-09-170.1.1
Source ↗ · Checked 2026-09-17CC0-1.0; package licenses are separate
Source ↗ · Checked 2026-09-17pypi
Source ↗ · Checked 2026-09-17cost-guard-mcp
Source ↗ · Checked 2026-09-190.3.0
Source ↗ · Checked 2026-09-19The publisher documents a single local Python stdio process spawned by an MCP host; this review did not connect to a warehouse or run a query.
Source ↗ · Checked 2026-09-19Supported warehouses use different credential models: Google application credentials, Snowflake user plus key or password, and Databricks token or OAuth client credentials; the mixed model is classified as unknown.
Source ↗ · Checked 2026-09-19Any compatible MCP client launches cost-guard-mcp over stdio and supplies the selected warehouse's connection settings through the host configuration.
Source ↗ · Checked 2026-09-19Publisher documentation reviewed; package and integration endpoint not executed or independently security-audited.
Source ↗ · Checked 2026-09-19