- Dataset
- 2026-10-27-catalog-ai-v2
- Ruleset
- b2e8889c66f7
- Scanned
- 2026-10-27
- Window
- February–September 15, 2026
UCCL is an efficient communication library for GPUs, covering collectives, P2P (e.g., KV cache transfer, RL weight transfer), and EP (e.g., GPU-driven)
01Result
uccl-project/uccl ↗uccl-project/uccl AI coding statistics: 6.5%
Adoption is broad. Many contributors are committing agent-authored work, which is what a real organisational rollout looks like.
02Drivers
What shapes this result
Mode, timing, and detected signatures show whether the result reflects broad adoption or one concentrated workflow.
Assisted or autonomous
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
| Month | Agent share | Considered commits |
|---|---|---|
| 2026-02 | 0% | 59 |
| 2026-03 | 3.1% | 65 |
| 2026-04 | 4.3% | 47 |
| 2026-05 | 7.3% | 41 |
| 2026-06 | 16.7% | 18 |
| 2026-07 | 26.7% | 15 |
| 2026-08 | 23.1% | 13 |
| 2026-09 | 0% | 2 |
Peak month: 26.7%. Bars are scaled to that peak, not to 100%.
Which agents
| Agent | Vendor | Commits | Share |
|---|---|---|---|
| Claude Code | Anthropic | 8 | 3.1% |
| Cursor Agent | Anysphere | 5 | 1.9% |
| GitHub Copilot | GitHub | 3 | 1.2% |
| OpenAI Codex | OpenAI | 2 | 0.8% |
03Evidence
Inspect the matches
Public commit samples let you inspect the signatures behind the aggregate result.
Matched commit samples
822a8fc0on GitHub[EP] size internode TMA sender buffer for hidden up to 8192 (#1020)Claude Code3e06114fon GitHub[P2P] Add Cambricon MLU (CNRT) support for cross-node P2P RDMA (#1011)Claude Code4312141con GitHub[Tooling] add pyproject.toml metadata and uv.lock (#434) (#1025)Claude Codea8c94e48on GitHub[CI] Skip CXI apt install when libfabric/hwloc are already present (#1041)Cursor Agent9ef7dfd4on GitHub[EP] Add UCCL_EP_DISABLE_CPU_AFFINITY to opt out of proxy thread CPU pinniCursor Agenta7176a99on GitHub[CI] Set NCCL_SOCKET_IFNAME for GH200 EP latency test (#1034)Cursor Agent7be91f17on GitHub[LITE]Reorganize lite to lite-ep and lite-collective (#962)GitHub Copilot8ac850bdon GitHub[UK] Unify buffer_id semantics across Transport/CCL/Python and harden OOB/GitHub Copilotbcd180e7on GitHub[EP] remove aten/torch depedency fromm uccl_ep (#784)GitHub Copilotc8b236f4on GitHub[EP] Split proxy CUDA device index from NIC local rank (#989)OpenAI Codex34bdf4f4on GitHub[EP] Support distinct HT barrier local ranks (#985)OpenAI 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 window | 260 |
| Merge commits removed | −0 |
| CI / dependency bots removed | −0 |
| Considered | 260 |
Bots are excluded from the denominator entirely. Agent attribution requires a literal supported signature; no code-style inference is used.
Continue exploring
Related coding-agent evidence
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 uccl-project/uccl