35.1%

of 19,133 commits were committed by an AI agent — a floor, not an estimate

35.1% agent · top contributor 39.9%100 contributors

org-wide

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

6,716

agent commits

12,417

human commits

100

agent contributors

39.9%

largest contributor

Assisted or autonomous

6713assisted

3autonomous

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

Jun26

Jul26

Aug26

Sep26

Peak month: 61.5%. Bars are scaled to that peak, not to 100%.

Which agents

Claude Code	Anthropic	6611	34.6%
Cursor Agent	Anysphere	96	0.5%
OpenAI Codex	OpenAI	12	0.1%
Devin	Cognition	4	0%

Verify it yourself

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

d584c350test(ui): give the client lanes a browser before the Story Gate timezone cClaude Code
2d03d182test(cloud): let the Worker placement test accept targeted placementClaude Code
f3a0173btest(scripts): build the receipt proof fixture with the continuity contracClaude Code
e908878ffix(plugin-elizacloud): retry total voice catalog outages (#23406)Cursor Agent
84bbd770fix(core): warm runtime probe cache off the boot critical pathCursor Agent
a30bac88fix(plugin-computeruse): reject PNG IHDR pixel bombs before inflate (#2230Cursor Agent
160a095arelease: promote media and voice lifecycle recovery (#28104)OpenAI Codex
d8ea6152fix(app): retire unreachable Android OTA checkers (#24265)OpenAI Codex
5ba741d7fix(app): preserve complete browser and OCR content (#24238)OpenAI Codex
b658d1e4feat(desktop): cloud-only settings panel with hardened Cloud contracts (#2Devin
f3ffaee9fix(desktop): handle binary responses in Electrobun HTTP bridge to fix TTSDevin
f43d944achore(electrobun): add cloud login E2E benchmark script (#22950)Devin

What was excluded

Commits in window	22,553
Merge commits removed	−3,294
CI / dependency bots removed	−126
Considered	19,133

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 elizaOS/eliza

Scan another repository

Try