Pulse
2026-09-18-v3.0
Dataset
2026-10-27-catalog-ai-v2
Ruleset
b2e8889c66f7
Scanned
2026-10-27
Window
February–September 15, 2026
Catalog tracking observation. This repository is measured with the same fixed window and ruleset, but it is not part of the published benchmark or its aggregate claims.

Lean 4 programming language and theorem prover

01Result

leanprover/lean4 ↗

leanprover/lean4 AI coding statistics: 20.7%

of 1,991 commits in February–September 15, 2026 were committed by an AI agent — a floor, not an estimate
20.7% agent · top contributor 26.9%22 contributors
org-wide

Adoption is broad. Many contributors are committing agent-authored work, which is what a real organisational rollout looks like.

412
agent commits
1,579
other considered commits
22
agent contributors
26.9%
largest contributor

02Drivers

What shapes this result

Mode, timing, and detected signatures show whether the result reflects broad adoption or one concentrated workflow.

Assisted or autonomous

412assisted
0autonomous

Assisted — a developer’s identity authored the commit and the agent added itself as co-author. Autonomous — a bot identity authored the commit without a human identity attached.

Agent share by month

Agent-attributed commit share by month
MonthAgent shareConsidered commits
2026-0216%382
2026-0334.7%352
2026-0427.8%216
2026-0516%188
2026-0614.7%218
2026-0714.1%298
2026-0816.2%228
2026-0925.7%109

Peak month: 34.7%. Bars are scaled to that peak, not to 100%.

Which agents

AgentVendorCommitsShare
Claude CodeAnthropic40320.2%
Cursor AgentAnysphere50.3%
GitHub CopilotGitHub40.2%

03Evidence

Inspect the matches

Public commit samples let you inspect the signatures behind the aggregate result.

Matched commit samples

  • ceed0430 on GitHubfeat: `ChoiceResolutionInfo` (#15070)Claude Code
  • 90dc0c50 on GitHubchore: recompile stdlib C files when `lean.h` changes (#15164)Claude Code
  • 368d1166 on GitHubfeat: bundle the `con-ron` external checker (#15153)Claude Code
  • f29e9e48 on GitHubfeat: support well-founded and state-dependent termination measures in SpeCursor Agent
  • 15adc1ca on GitHubchore: correct vcgen bench toolchain and move IMP test to elab (#14504)Cursor Agent
  • f82e306c on GitHubfix: activate CompleteLattice→CCPO under open Std.Internal.Do (#14502)Cursor Agent
  • 91513604 on GitHubfix: preserve symbol hover for `fun_induction` function target (#13678)GitHub Copilot
  • 6ebe573c on GitHubfix: kernel: move level parameter count and thm-is-prop checks for robustnGitHub Copilot
  • 73640d37 on GitHubfix: preserve `@[implicit_reducible]` for WF-recursive definitions (#12776GitHub Copilot

04Method

What entered the denominator

The result begins with commit history, removes merges and routine automation, and reports literal agent signatures as a lower bound.

Exclusion ledger

Commits in window2,001
Merge commits removed−0
CI / dependency bots removed−10
Considered1,991

Bots are excluded from the denominator entirely. Agent attribution requires a literal supported signature; no code-style inference is used.

Continue exploring

Claude Code commitsBrowse repositories with detected Claude Code signatures →GitHub Copilot commitsBrowse repositories with detected GitHub Copilot signatures →Cursor Agent commitsBrowse repositories with detected Cursor Agent signatures →GitHub AI commit analyzerInspect another repository with the same evidence model →
Reproduce this observation

Run the same detector locally.

Deterministic signature matching; no code-style inference. Squash merges can remove trailers, so this result remains a measured floor.

npx qmmit-cli leanprover/lean4

Scan another repository

Try