12.3%

of 7,893 commits were committed by an AI agent — a floor, not an estimate

12.3% agent · top contributor 36.3%45 contributors

org-wide

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

968

agent commits

6,925

human commits

agent contributors

36.3%

largest contributor

Assisted or autonomous

948assisted

20autonomous

Assisted — a developer’s identity authored the commit and the agent added itself as co-author. The agent ran git commit during a working session, with a person there. Autonomous — a bot identity is the author, no human in the loop. Across the whole index, 93.7% of agent commits are assisted.

Agent share by month

Feb26

Mar26

Apr26

May26

Jun26

Jul26

Aug26

Sep26

Peak month: 42.2%. Bars are scaled to that peak, not to 100%.

Which agents

Claude Code	Anthropic	838	10.6%
Cursor Agent	Anysphere	48	0.6%
Devin	Cognition	44	0.6%
GitHub Copilot	GitHub	33	0.4%
Gemini CLI	Google	6	0.1%
OpenAI Codex	OpenAI	2	0%

Verify it yourself

Sample commits that matched. Every one links to GitHub — check the trailers.

1c29c442fix(review): keep the literal-glob test on Windows instead of gating it ofClaude Code
fbf71588chore(ci): budget the editor-handoff pause sleep (#6149)Claude Code
e8a22655docs(runtime-api): describe the command catalog the merged routes actuallyClaude Code
f76252d0fix(providers): Ollama default comes from the live local catalog, not a hoCursor Agent
55b53b6cfeat(providers): Concentrate as a first-class opt-in BYOK Responses gatewaCursor Agent
72baeeb3feat(cli): cloud-dispatch remote runner — sandbox to forge PR (#5712)Cursor Agent
060711b2feat(runtime-api): native-client routes for jobs, context, secrets, git, dDevin
7c74ecf0tui: bound the turn-end foreground-child join so a stuck child cannot withDevin
1c21a6bcMake a goal's token budget an optional hard stop (#6013)Devin
b36e2c2btui: 0.9.12 defect pass — launch clock, dock, /mcp room, automations room,GitHub Copilot
ba386a0cskills: three-tier ambient index budgeted from the context window; names nGitHub Copilot
2742abfftui: launch card runs nothing on a reflexive Enter; every launch flow has GitHub Copilot
29383349Update crates/tui/src/localization.rsGemini CLI
2c8f7e2aUpdate docs/INSTALL.mdGemini CLI
e439f371Update docs/INSTALL.mdGemini CLI
c97c3a7aSummarize Cargo failures in tool metadata (#1973)OpenAI Codex
16728360Expose apply_patch preflight metadata (#1971)OpenAI Codex

What was excluded

Commits in window	9,731
Merge commits removed	−1,727
CI / dependency bots removed	−111
Considered	7,893

Bots are excluded from the denominator entirely. Counting dependabot as “AI-written code” is the fastest way to produce a meaningless number.

How this was measured

Measured over February–August 2026, cloned directly. Detection is deterministic string matching against signatures agents write about themselves — no heuristics, no style analysis. Merge commits and CI bots are excluded from the denominator. The figure is a floor: squash merges strip trailers, so real usage is higher. Contributor counts are aggregate — no individual is named.

Run it yourself: npx qmmit-cli Hmbown/Codewhale

Scan another repository

Try