Your dashboard says you’re shipping faster. What is it costing in review?
Point this at a public repository. It reads the recently merged pull requests, separates the ones carrying agent authorship from the rest, and compares what each cost in human review rounds, rework, and time to merge.
Public repositories only. Nothing is installed and no account is created. Try , , .
Across 392 published reports covering 38,457 merged pull requests, the median team’s attributed work takes 0.6× as long to merge as the rest.
What it can and cannot see
Attribution comes only from marks a tool leaves on a commit — a co-author trailer, an agent commit identity, an agent bot account. Tools that just complete code inline leave nothing, so the attributed share is a floor and the comparison column is “the rest”, never “human-written”. Nothing is inferred from diff shape or code style. The full method, and where it’s wrong
Private repository?
This form only reads public repos. For a private one, run the same read locally — it goes through your own GitHub credentials, installs nothing, and sends nothing to us.
npx @ambera/review-taxIt also runs as an MCP server, so you can ask a coding agent directly: claude mcp add review-tax -- npx -y @ambera/review-tax --mcp
Want this continuously, with your team’s estimates calibrated against it? That’s Ambera Forge · See what others have read