Pulse
2026-09-18-v3.0
Dataset
2026-10-27-catalog-ai-v2
Ruleset
b2e8889c66f7
Scanned
2026-10-27
Window
February–September 15, 2026
Catalog tracking observation. This repository is measured with the same fixed window and ruleset, but it is not part of the published benchmark or its aggregate claims.

The HIP Environment and ROCm Kit - A lightweight open source build system for HIP and ROCm

ROCm/TheRock ↗

ROCm/TheRock AI coding statistics: 28.1%

of 1,792 commits in February–September 15, 2026 were committed by an AI agent — a floor, not an estimate
28.1% agent · top contributor 22.9%77 contributors
org-wide

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

503
agent commits
1,289
other considered commits
77
agent contributors
22.9%
largest contributor

What shapes this result

Mode, timing, and detected signatures show whether the result reflects broad adoption or one concentrated workflow.

Assisted or autonomous

503assisted
0autonomous

Assisted — a developer’s identity authored the commit and the agent added itself as co-author. Autonomous — a bot identity authored the commit without a human identity attached.

Agent share by month

Agent-attributed commit share by month
MonthAgent shareConsidered commits
2026-0217.6%233
2026-0325%276
2026-0425.9%255
2026-0531.6%190
2026-0636%272
2026-0724.9%241
2026-0830.6%219
2026-0939.6%106

Peak month: 39.6%. Bars are scaled to that peak, not to 100%.

Which agents

AgentVendorCommitsShare
Claude CodeAnthropic43424.2%
Cursor AgentAnysphere563.1%
GitHub CopilotGitHub130.7%
OpenAI CodexOpenAI70.4%

Inspect the matches

Public commit samples let you inspect the signatures behind the aggregate result.

Matched commit samples

  • 1d8c858b on GitHubci(rocgdb): narrow corefile test job to runtime-core.exp only (#8215)Claude Code
  • eec5e937 on GitHubfeat(ci): disable rocminfo sanity check for gfx125X-dcgpu (#8225)Claude Code
  • ff2bcd72 on GitHubfix(compiler): improve compiler bump workflow (#8221)Claude Code
  • 5cbabe71 on GitHubfix(ci): restore rocm-systems DVC pull for external rocm-libraries builds Cursor Agent
  • 21e69070 on GitHubAdd test filter standardisation for rccl (#5469)Cursor Agent
  • 42097d0a on GitHubfix(ci): skip the libraries baseline gate instead of failing it (#8161)Cursor Agent
  • 4665f854 on GitHubdocs(windows): bump winget Windows 11 SDK example to 26100 (#8112)GitHub Copilot
  • 648a501e on GitHubfix(ci): guard Unix-only fcntl import in Windows driver/GPU sanity check (GitHub Copilot
  • 851984e8 on GitHub(feat) (AICV-246) Build amd-mesa/libva sysdep on Windows (#7279)GitHub Copilot
  • ca66f476 on GitHub[release] Support BKC release types in release workflows (#7392)OpenAI Codex
  • c6fac2c1 on GitHub[packaging] Add versioning for BKC release type (#7390)OpenAI Codex
  • 836c7121 on GitHub[packaging] Refine package version tests and documentation (#7389)OpenAI Codex

What entered the denominator

The result begins with commit history, removes merges and routine automation, and reports literal agent signatures as a lower bound.

Exclusion ledger

Commits in window2,007
Merge commits removed−0
CI / dependency bots removed−215
Considered1,792

Bots are excluded from the denominator entirely. Agent attribution requires a literal supported signature; no code-style inference is used.

Reproduce this observation

Run the same detector locally.

Deterministic signature matching; no code-style inference. Squash merges can remove trailers, so this result remains a measured floor.

npx qmmit-cli ROCm/TheRock

Scan another repository