← the index
topoteretes/cognee ↗
29.6%
of 3,419 commits were committed by an AI agent — a floor, not an estimate
29.6% agent · top contributor 44%34 contributors
org-wide

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

1,012
agent commits
2,407
human commits
34
agent contributors
44%
largest contributor

Assisted or autonomous

1011assisted
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: 48.2%. Bars are scaled to that peak, not to 100%.

Which agents

Claude CodeAnthropic99729.2%
Cursor AgentAnysphere140.4%
GitHub CopilotGitHub30.1%

Verify it yourself

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

  • a9b5575bdocs: Drop root-endpoint claim from health blurb (RES-21)Claude Code
  • ffec361edocs: Fix three inaccurate tag blurbs (RES-21)Claude Code
  • adafe1defix: Read EMBEDDING_API_BASE as endpoint alias (SDK-539) (#4882)Claude Code
  • 86553196fix(storage): Write the graph before the vectorsCursor Agent
  • 1f2e3098fix(update): Keep the original owner on full-rebuild fallbackCursor Agent
  • eb801d54fix(update): Authorize incremental update by dataset ACLCursor Agent
  • ab97e48ePotential fix for pull request finding 'CodeQL / Workflow does not containGitHub Copilot
  • 0a264bc3Fix LanceDB test process hang by closing vector engineGitHub Copilot
  • 607c55d3Potential fix for pull request finding 'CodeQL / Information exposure throGitHub Copilot

What was excluded

Commits in window5,011
Merge commits removed−1,587
CI / dependency bots removed−5
Considered3,419

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 topoteretes/cognee

Scan another repository

Try