Coverage MCP
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Coverage MCPshow me coverage changes since last snapshot"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Coverage MCP
Find missing coverage, inspect the returned locations with rg, run your
existing test command, and compare the reports. Coverage MCP reads LLVM JSON,
LCOV, coverage.py JSON, Istanbul JSON, Cobertura, JaCoCo, and Go coverprofiles.
Development status: this checkout is 0.16.0, schema 18. It replaces the old
API with two read-only tools: coverage_gaps and coverage_compare. The local
runtime and marketplace update have been tested; publishing them is a separate
release step. See upgrading from 0.15.
Install
For Codex, install the testing plugin from
codegen-marketplace.
The plugin pins the runtime version and starts the shared daemon. The published
plugin may still provide the earlier contract until 0.16 is released.
For contributors testing this checkout, Rust 1.85.1 and a C++ build toolchain are required for the bundled DuckDB build:
cargo install --path . --locked --bin coverage-mcp
coverage-mcp --versionRelated MCP server: SNAP MCP Server
Try a report
After installation, run this in a project directory:
mkdir -p target
printf 'TN:\nSF:example.rs\nDA:1,1\nDA:2,0\nend_of_record\n' > target/example.lcov
coverage-mcp gaps --measurement target/example.lcovThe result identifies line 2 as missing. Source identity is unverified and test
status is unknown; the example does not require an example.rs source file.
For real measurements, preserve the baseline before running tests:
cp target/coverage.json target/baseline.json
# Run only the new cases with the repository's existing coverage command.
coverage-mcp gaps --measurement target/coverage.json --query parse_header
coverage-mcp compare --scope incremental --baseline target/baseline.json --measurement target/coverage.jsonIncremental mode reports new coverage and the baseline-plus-batch total. Pass
--previous REPORT for earlier batches; see incremental use.
Report-file CLI queries need no server, database, or import. Each answer shows
a few source locations and omitted counts. Selected-test absences are not
regressions; source/build receipts are needed for verified comparison claims.
Use --json for structured output or --help for command options.
To use MCP and saved history, run coverage-mcp connect --repo /absolute/project.
It starts or reuses the loopback daemon; the dashboard is then available at
http://127.0.0.1:59471/.
Guides
For bugs, see support. Report vulnerabilities through the security policy. Licensed under MIT.
This server cannot be deployed
Maintenance
Related MCP Connectors
AI Reasoning Cache & Consensus Layer with 11 MCP tools via Streamable HTTP.
MCP server for visual regression testing: triage a PR's UI diffs from your coding agent.
MCP server providing access to the Scorecard API to evaluate and optimize LLM systems.
Repository knowledge graph MCP server for codebase understanding and debugging.
Related MCP Servers
- AlicenseAqualityFmaintenanceMCP server for Codecov that provides tools to get commit coverage totals and prompts to suggest tests to write.138 npm6ISC
- FlicenseNot gradedqualityDmaintenanceA Model Context Protocol server that parses code, documents, data, and config files into structured database snapshots for targeted AI retrieval, enabling code analysis and project understanding via LLM interactions.1-
- FlicenseNot gradedqualityAmaintenanceLocal MCP server that lets your AI coding agent query its own cross-tool project history - file/command freshness, past test failures, cost & token spend, cache status, and session handoff - over stdio, 100% local, no telemetry.46-
- AlicenseNot gradedqualityDmaintenanceMCP server that provides instant, token-efficient test coverage data for any programming language, allowing agents to query coverage summaries and track session progress.27 npm41MIT