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.

Framework for evaluating and improving agents

01Result

harbor-framework/harbor ↗

harbor-framework/harbor AI coding statistics: 23.9%

of 1,176 commits in February–September 15, 2026 were committed by an AI agent — a floor, not an estimate
23.9% agent · top contributor 24.6%107 contributors
org-wide

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

281
agent commits
895
other considered commits
107
agent contributors
24.6%
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

280assisted
1autonomous

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
2026-010%1
2026-0219.2%99
2026-0327.7%148
2026-0425.1%179
2026-0531.7%101
2026-0624.6%203
2026-0729.9%177
2026-0819.2%177
2026-097.7%91

Peak month: 31.7%. Bars are scaled to that peak, not to 100%.

Which agents

AgentVendorCommitsShare
Claude CodeAnthropic16914.4%
Cursor AgentAnysphere827%
DevinCognition494.2%
GitHub CopilotGitHub90.8%
Other / generic agentVarious70.6%
Gemini CLIGoogle40.3%

03Evidence

Inspect the matches

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

Matched commit samples

  • 4797595c on GitHubfix(antigravity-sdk): take token totals from the SDK session aggregate (#3Claude Code
  • 5c364a53 on GitHubfeat(swe-agent): expose completion_kwargs (#3048)Claude Code
  • e348ba30 on GitHubdocs: drop stale pinned version from README citation block (#3027)Claude Code
  • 4008e2df on GitHubfix(multi-step): don't leak prior step's /tests and /logs/verifier into thCursor Agent
  • 3e39b472 on GitHubfix(codex): accept none reasoning effort for luna (#3087)Cursor Agent
  • ac476798 on GitHubDeclare and validate built-in agent options (#3049)Cursor Agent
  • b5c76f1f on GitHubAdd Muse Code installed agent integration (#3043)Devin
  • 8552e6be on GitHubDefault `harbor auth login` to the manual flow in SSH sessions (#2981)Devin
  • 283ef4ee on GitHubfeat: add Podman as a ContainerRuntime behind DockerEnvironment (#2872)Devin
  • 61095f40 on GitHubfix(docker): stream stdout without a line-length limit (#2985)GitHub Copilot
  • bc559442 on GitHubFix Windows cross-drive artifact paths (#2878)GitHub Copilot
  • e68e5b49 on GitHubfix(copilot-cli): map streamable-http MCP transport to http (#2257)GitHub Copilot
  • 8cf5fdfb on GitHubfix(openhands-sdk): use uv to pin Python >=3.12 for SDK venv (#2041)Other / generic agent
  • b60341cb on GitHub[Ready for Review - bot review] HLE Adapter (#878)Other / generic agent
  • 3a7e4aa6 on GitHub[Ready for Review] Adapter: ML dev bench (#157)Other / generic agent
  • 5fcf07f5 on GitHub[Ready for Review - human review] Research code bench (#477)Gemini CLI
  • 006e127c on GitHub[Ready for Review - human review] Adapter: Omni-Math (#406)Gemini CLI
  • 593799e4 on GitHub[Ready for Review] Update MLGym-Bench Adapter for GPU and Continuous MetriGemini CLI

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 window1,216
Merge commits removed−40
CI / dependency bots removed−0
Considered1,176

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 →GitHub Copilot commitsBrowse repositories with detected GitHub Copilot signatures →Cursor Agent commitsBrowse repositories with detected Cursor Agent signatures →Devin commitsBrowse repositories with detected Devin signatures →Gemini CLI commitsBrowse repositories with detected Gemini CLI 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 harbor-framework/harbor

Scan another repository

Try