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.

A modular, primitive-first, python-first PyTorch library for Reinforcement Learning.

01Result

pytorch/rl ↗

pytorch/rl AI coding statistics: 17.5%

of 650 commits in February–September 15, 2026 were committed by an AI agent — a floor, not an estimate
17.5% agent · top contributor 92.1%6 contributors
individual

One person is driving this. The headline number reflects a single contributor's workflow, not a project-wide shift.

114
agent commits
536
other considered commits
6
agent contributors
92.1%
largest contributor

02Drivers

What shapes this result

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

Assisted or autonomous

114assisted
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
2026-0256.8%81
2026-036.7%15
2026-041.7%60
2026-052.7%74
2026-0612.2%98
2026-0720.5%88
2026-088.4%107
2026-0919.7%127

Peak month: 56.8%. Bars are scaled to that peak, not to 100%.

Which agents

AgentVendorCommitsShare
Claude CodeAnthropic649.8%
Cursor AgentAnysphere497.5%
OpenAI CodexOpenAI10.2%

03Evidence

Inspect the matches

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

Matched commit samples

  • 2d258feb on GitHub[Feature] Add resume= to the standalone SAC, TD3 and DDPG recipes (#4385)Claude Code
  • e68c8f33 on GitHub[Feature] Add resume= to the Hydra trainer recipes (#4384)Claude Code
  • 5d69706a on GitHub[BugFix] Reject W&B logger state from another run and add Checkpoint.read_Claude Code
  • 2d2b65a2 on GitHub[Doc] Migrate shifted=True callers to legacy/compact + docstring polishCursor Agent
  • c6d118b9 on GitHub[Performance] LSTM/GRU scan: canonical strides + cuDNN flat-storage clonesCursor Agent
  • 4e2e787b on GitHub[Feature] Add SGLang backend support to GRPOCursor Agent
  • 54f6d9cc on GitHub[Feature] Add MuJoCo WASM notebook backendOpenAI Codex

04Method

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 window671
Merge commits removed−0
CI / dependency bots removed−21
Considered650

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

Continue exploring

Claude Code commitsBrowse repositories with detected Claude Code signatures →Cursor Agent commitsBrowse repositories with detected Cursor Agent signatures →OpenAI Codex commitsBrowse repositories with detected OpenAI Codex signatures →Python AI coding statisticsCompare evidence across Python repositories →GitHub AI commit analyzerInspect another repository with the same evidence model →
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 pytorch/rl

Scan another repository

Try