← the index
roboflow/supervision ↗
66.4%
of 399 commits were committed by an AI agent — a floor, not an estimate
66.4% agent · top contributor 34.7%66 contributors
org-wide

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

265
agent commits
134
human commits
66
agent contributors
34.7%
largest contributor

Assisted or autonomous

261assisted
4autonomous

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: 85.3%. Bars are scaled to that peak, not to 100%.

Which agents

Claude CodeAnthropic15338.3%
GitHub CopilotGitHub12130.3%
OpenAI CodexOpenAI8020.1%
Cursor AgentAnysphere30.8%

Verify it yourself

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

  • c6f197d3fix(dataset): size EXIF-oriented images the way they load (#2577)Claude Code
  • 3bb7a41ffix(detection): load v5 instance results returned as binary maps (#2576)Claude Code
  • 7c9bdb7ffix(key_points): place Inference key points by their skeleton index (#2575Claude Code
  • d4132ce3test: fix ultralytics regression fixture lengthGitHub Copilot
  • 16482578fix(utils): stop image resizing from deriving a zero-sized target (#2564)GitHub Copilot
  • 7e9a98e6docs: convert fenced examples to doctests in detection/core.py (#2558)GitHub Copilot
  • 3bb7a41ffix(detection): load v5 instance results returned as binary maps (#2576)OpenAI Codex
  • 7c9bdb7ffix(key_points): place Inference key points by their skeleton index (#2575OpenAI Codex
  • bc5cc906feat(metrics): add `MetricResult` ABC and aggregate comparison functions (OpenAI Codex
  • 9837c178feat(vlm): add Gemini 3.5 Flash parsing support (#2449)Cursor Agent
  • caf06052refactor: enrich `KeyPoints` with visibility mask, split confidence fieldsCursor Agent
  • 5a8e211aDocs/api design principles contributing (#2292)Cursor Agent

What was excluded

Commits in window477
Merge commits removed−0
CI / dependency bots removed−78
Considered399

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 roboflow/supervision

Scan another repository

Try