An MCP server that records AI coding agent attempts, runs your test suite against each in isolated git worktrees, and reports pass/fail verdicts to help you choose the fix that actually works.
An MCP server that lets AI agents debug Python code by running scripts, setting breakpoints, stepping through execution, inspecting live variable values, and editing or reverting files to verify fixes. It relies only on Python's standard sys.settrace, requiring no debugpy, DAP, pdb, or PTY.
This MCP server gives AI coding agents structured, scoped access to GitHub issues, a codebase, its docs, and its test suite, enabling issue triage, code search, test execution, and failure analysis in one workflow. It provides deterministic tools for fetching and analyzing issues, searching code, reading documentation, running tests, and summarizing failures.