3.1%

of 25,203 commits were committed by an AI agent — a floor, not an estimate

3.1% agent · top contributor 6.4%333 contributors

org-wide

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

780

agent commits

24,423

human commits

333

agent contributors

6.4%

largest contributor

Assisted or autonomous

779assisted

1autonomous

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: 3.7%. Bars are scaled to that peak, not to 100%.

Which agents

Claude Code	Anthropic	679	2.7%
Cursor Agent	Anysphere	72	0.3%
GitHub Copilot	GitHub	27	0.1%
OpenAI Codex	OpenAI	4	0%
Other / generic agent	Various	1	0%

Verify it yourself

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

d47fe28ffix(backup): import SQLite databases without replacing the live inodeClaude Code
8381d4facompression: emit periodic client-visible status during compactionClaude Code
e73257ccfeat(cli): filter MCP server spawning by -t/--toolsets flag + fix orphan sClaude Code
7f7df1cefix(state): refuse SessionDB open and writes on a deleted WAL generationCursor Agent
622883bafix(agent): accept marker-only finish_reason after stream supersessionCursor Agent
ce7f8058fix(proxy): append SSE [DONE] when Nous streams omit the sentinelCursor Agent
0b50c8e4ci: wrap uv python install in retry action on OS test laneGitHub Copilot
141d5db9fix(cron): scope in_channel thread_id clear to the live-send/seed gateGitHub Copilot
5f2fdf66feat: per-model compression threshold overrides (v2, rebased on main)GitHub Copilot
b595fcd5fix(desktop): keep the Linux HUD clickable and recoverableOpenAI Codex
cfb55de5Update Stripe Projects skill docs (#48673)OpenAI Codex
17588521fix(docs): align fallback provider config examplesOpenAI Codex
7ebebfbbHarden Skills Guard multi-word prompt patterns (#26852)Other / generic agent

What was excluded

Commits in window	27,511
Merge commits removed	−2,004
CI / dependency bots removed	−304
Considered	25,203

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 NousResearch/hermes-agent

Scan another repository

Try