AI Dev Impact Lab JA
← Monthly reports · REPORT 2026-05 · published 2026-06-02
MONTHLY REPORT

May 2026 AI Development Impact Report

AI development on GitHub in 2026-05. In a random sample of 1,784 repositories, 8.0% were AI-built and 4.6% used AI, alongside this month's coding-tool adoption.

3,448
AI repos observed
1,265
tool signals
453
solo dev
518
web launches
AI SHARE
8.0%
AI-using 4.6%
Sample n 1,784
AI-built share over time
🥇 TOP TOOL
Claude Code
711
repos
57.8%
share

ADOPTION AI coding tool adoption — May 2026

CONTEXT Global GitHub scale (GH Archive)

4,981,000
new repos
10,561,034
create events
88,682,757
push events
248,401
forks
986,355
watch events

Note: GH Archive's fork and watch/star coverage degrades sharply from late 2025. Values are shown as the source publishes them.

ANALYSIS Analysis & commentary

Correction, August 2026: the text below reasons from a "composite score" that counted how many repositories this site managed to collect, not activity on GitHub. The fall through June and July 2026 in particular was a defect in our collection, not a real decline. The metric has been retired in favour of the sampled share shown above. The body is left as originally published, for the record.

Executive summary

Observed facts: For 2026-05 the composite score is 1,402.45 (unit: score). The input also contains composite_score_mom = 1,213.77 and composite_score_yoy = 1,213.77 (unit: score). There are 3,448 AI-related repositories and 1,265 repositories carrying AI-tool signals. The reported continuity_rate is 0 (confidence: low). Global totals include 11,747,368 active repositories.

Interpretation: The composite score shows a sharp numeric elevation compared with the long run of 106.75 in prior months (listed in the recent history), indicating concentrated activity or scoring changes rather than broad growth across the entire index. The low-confidence continuity_rate of 0 should be treated cautiously; it may reflect a measurement limitation rather than complete lack of ongoing work.

Key movements

Observed facts: Tool signals are concentrated under a few labels: claude_md appears in 727 repositories, agents_md in 434 repositories, cursorrules in 51 repositories, and copilot_instructions in 36 repositories. Top individual repositories by stars include nexu-io/open-design with 35,105 stars (AI score 60), 1weiho/open-slide with 2,778 stars (AI score 45), and WenyuChiou/awesome-agentic-ai-zh with 516 stars (AI score 100). The dataset lists 453 solo development candidates and 518 web launch candidates.

Interpretation: The prevalence of claude_md and agents_md signals suggests attention on agentic and conversational tooling flows. High-star TypeScript and Python projects indicate mainstream developer interest concentrated in a small number of visible projects rather than widespread, diffuse adoption. The solo and web candidate counts point to potential near-term public launches.

Global context

Observed facts: Platform-wide activity includes 10,561,034 create events, 24,8401 fork events, 88,682,757 push events, and 986,355 watch events. Total public repositories created are 4,981,000 and total active repositories are 11,747,368.

Interpretation: AI-related repositories (3,448) represent about 0.029% of active repositories (3,448 / 11,747,368 ≈ 0.029%). Repositories carrying AI-tool signals (1,265) represent about 0.011% of active repositories (1,265 / 11,747,368 ≈ 0.011%). These fractions indicate that current AI-labeled activity is a small portion of overall platform activity in absolute terms, even where intensity metrics like the composite score are elevated.

Outlook

Observed facts: The composite score in 2026-05 is 1,402.45 and recent monthly history shows stable 106.75 for many months before this value. Key signals (claude_md 727, agents_md 434) and candidate counts (453 solo, 518 web) are present in the data.

Interpretation and recommended monitoring: Continue to monitor the composite score and the counts of AI-tool-signal repositories to determine whether the current score elevation persists or reverts. Investigate the continuity_rate measurement (reported as 0 with low confidence) to clarify whether it reflects a data artifact. Track top tool signals and the pool of solo/web candidates for early indicators of broader adoption or public launches.

REPOS Top repos in May 2026 (AI score 40+ / by stars)

nexu-io/open-design

🎨 Best DeepSeek Harness Design Plugin. The open-source Claude Design alternative. 🖥️ Local-first desktop app. 🖼️ Your coding agent becomes the design engine: prototypes, landing pages, dashboards, slides, images & video — real files, HTML/PDF/PPTX/MP4 export. 🤖 Claude Code / Codex / Cursor / DeepSeek Harness / OpenCode & 20+ CLIs via BYOK.

TypeScript 35,105 AI 60 2 sig
1weiho/open-slide

A slide framework built for agents.

TypeScript 2,778 AI 45 2 sig
WenyuChiou/awesome-agentic-ai-zh

A trilingual (繁中 / English / 简中) learning roadmap for agentic AI: from LLM basics to multi-agent systems, with 240+ curated resources and hands-on examples. 中文 AI agent 學習地圖。

Python 516 AI 100
lonr-6/cc-desktop-switch

Lightweight desktop tool for configuring DeepSeek, Kimi, Zhipu GLM and Bailian providers in Claude Desktop.

Python 347 AI 45
99xAgency/GodModeSkill

Multi-LLM cross-review workflow for Claude Code. /work orchestrates plan/implement/bug-fix with 3 different model families voting on every gate.

Shell 215 AI 45
thinkpixelIab/polymarket-ai-trading

Polymarket prediction markets AI trading paper trading OpenAI GPT CLOB Kelly mean reversion SQLite Node Express Docker Render Vercel dashboard algorithmic trading quant research forecasting crypto

HTML 135 AI 90
MobiusQuant/OpenMobius-skill

ICT/SMC trading-knowledge skill for AI coding agents (Claude Code / Codex / OpenClaw / Hermes)

Python 89 AI 45
akitaonrails/ai-usagebar

Rust-based waybar widget to monitor status of Claude, GPT, GLM, OpenRouter plans/credits - inspired by claudebar/codexbar

Rust 76 AI 75 1 sig
whchoi98/claude-code-workshop

claude-code-workshop

HTML 67 AI 75
EliasOulkadi/shokunin

職人 Shokunin 62 AI agent skills for OpenCode, Claude Code, Cursor, Windsurf. ChromaDB memory, MCP servers, declarative self-updates. Multi-model, open source, zero cost.

HTML 63 AI 70
myICOR/myPKA

AI-powered Personal Knowledge Assistance in a folder. Built on the ICOR methodology. Plain markdown. Any LLM. Yours forever.

TypeScript 56 AI 70 2 sig
ProsecutorAqueduct/Claude-Pro-Crack-2026

⭐️ Claude Pro Download | AI Assistant Setup Installer v1.0 | Activation Key Keygen | Pre-Activated Full Version | Latest Build Pro 2026 | Get Desktop Working Windows 10/11 PC | Premium Loader Mod | Serial License Key | AI Chatbot App Software ⭐️

TypeScript 51 AI 80

NEW New AI repos created in May 2026

WenyuChiou/awesome-agentic-ai-zh

A trilingual (繁中 / English / 简中) learning roadmap for agentic AI: from LLM basics to multi-agent systems, with 240+ curated resources and hands-on examples. 中文 AI agent 學習地圖。

Python 516 AI 100 NEW
kjfx/AI

Claude Code 零基础入门教程,claude code 大陆如何使用教程

29 AI 100 NEW
waitdeadai/llm-dark-patterns

Umbrella for the LLM Dark Patterns Hooks suite — single-purpose Claude Code Stop hooks that suppress sycophancy, paternalism, false-success, permission-loops, training-cutoff confidence at the textual boundary.

Shell 10 AI 100 NEW
charlieviettq/awesome-agent-skill

Curated skill pack for LLM agents in engineer and science workflow (Cursor & Claude ready).

Python 6 AI 100 NEW
sebastianbreguel/agent-complexity-optimizer

Scan any codebase for O(n²) hotspots, N+1 queries, and complexity issues — works with 13+ AI coding agents

Python 3 AI 100 NEW
buraksenyurt/rust-llm-wiki

Andrej Karpathy'nin LLM Wiki önerisini Copilot ile denediğim çalışma alanı.

2 AI 100 NEW
japananh/aimonitor

Multi-account Claude Code usage monitor & silent account switcher, plus a Slack/ClickUp MCP server — macOS menu bar + CLI. Local-first, no telemetry.

Go 2 AI 100 NEW
TanChuping/Claude-voyager

GPL-3.0 browser extension for Claude.ai: timeline navigation, text pins, folders, favorites, prompt vault, and one-click conversation export.

TypeScript 2 AI 100 NEW

EXPLORE Dig deeper

📅

May 2026 raw snapshot

Top 20 repos, all monthly metrics, GH Archive globals

⌨️

AI coding tool ranking

Claude Code, Cursor, Codex, Copilot, Windsurf, Aider, Cline

📊

All 13 rankings

languages, LLMs, frameworks, DBs, models, embeddings, agents

Next → June 2026 AI Development Impact Report

Aggregated from snapshots of AI-relevant repos (AI score ≥ 40) for this month. methodology