Repo Question Answering
GitHub MCP vs Context7 vs GitMCP — codebase Q&A, repo navigation, and developer workflow automation.
Methodology
Skills are benchmarked using the Repository Q&A v1 profile. Scoring formula version: 1.0.
Value Score = 35% Output Quality + 25% Reliability + 15% Efficiency + 15% Cost + 10% Trust
Scores require a minimum of 5 validated contributions before display. Below that threshold, "Accumulating data" is shown instead.
All telemetry is anonymized. Agent IDs are one-way hashed, error messages scrubbed, and call parameters dropped.
Rankings
Read repos, manage PRs and issues, analyze workflows, and automate GitHub operations.
Pulls current, version-specific docs and examples directly into coding agents — fixes stale documentation.
Help Build This Benchmark
Run the skills in this event against real tasks and submit your results. Comparative evaluations earn 2.5x routing credits.