1.6%

of 4,403 commits were committed by an AI agent — a floor, not an estimate

1.6% agent · top contributor 19.4%37 contributors

org-wide

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

agent commits

4,331

human commits

agent contributors

19.4%

largest contributor

Assisted or autonomous

72assisted

0autonomous

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: 4%. Bars are scaled to that peak, not to 100%.

Which agents

Claude Code	Anthropic	49	1.1%
Gemini CLI	Google	18	0.4%
Cursor Agent	Anysphere	4	0.1%
GitHub Copilot	GitHub	2	0%
OpenAI Codex	OpenAI	1	0%

Verify it yourself

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

ba34b8f7Studio: spill weights with -ot instead of layers, so the KV cache stays onClaude Code
055bfd50Show all local models in the unsloth chat picker (#9435)Claude Code
86a89930Studio: keep unrunnable speech GGUFs out of the model pickers (#9433)Claude Code
c356427fGuard Windows ROCm torchao override skip (#6837)Gemini CLI
01f7e149Fix Studio custom folders on Linux external drives (#6799)Gemini CLI
22cd26f7feat: Implementation of the Portuguese (Brazil) language and VRAM/RAM moniGemini CLI
b38490f0fix(AMD): retry ROCm llama-server with bundled HIP on HIP/ROCR mismatch (#Cursor Agent
7367955bfix: replace deprecated transformers attention mask imports (#6880)Cursor Agent
515abca8fix(studio): adopt server-loaded model before chat auto-load (#5900)Cursor Agent
11469a60(feat) Add project names to studio training runs (#6512)GitHub Copilot
9a0d6f80studio: API external provider support for chat (OpenAI, Mistral, Gemini, CGitHub Copilot
34e22efbStudio: fix the hub download rate estimator and bound long ETAs (#7679)OpenAI Codex

What was excluded

Commits in window	4,746
Merge commits removed	−288
CI / dependency bots removed	−55
Considered	4,403

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 unslothai/unsloth

Scan another repository

Try