Anthropic Just Took GitHub’s #1 Spot With Its Own Repo — Dev Trends, Week of May 6–12



WEEKLY DEV TRENDS · 2026.05.06—05.12

Anthropic Just Took GitHub’s #1 Spot With Its Own Repo — Dev Trends, Week of May 6–12

This week’s GitHub trending #1 was a financial-services repo Anthropic published itself. It pulled 13,000 stars in seven days. The theme of the week: where and how do you run an AI agent?

Anthropic 13K★ Local AI NPM Supply-Chain Hit Cloudflare -20%

01 · MAINThis week’s #1 was Anthropic itself

anthropics/financial-services · STARS THIS WEEK
0 ★
(at launch)
+13,176 ★

21,500★ TOTAL · GITHUB TRENDING #1

An official guide repo bundling Claude system prompts, MCP references, and evaluation datasets for the financial domain.

Anthropic’s official GitHub org (anthropics/) shipped a full Claude guide for financial services. System prompt templates, MCP server references, and evaluation datasets all live in one repo, which the community seems to have picked up as the canonical example of “how to build an AI agent”. 13K stars in a week = trending #1.

02 · RANKGitHub weekly Top 5 — half of it is agent infra

1

anthropics/financial-services

PYTHON · +13,176 ★ / WEEK · TOTAL 21,500★

Claude guide for the financial domain — system prompts, MCP, eval sets. Anthropic itself takes #1.

2

Hmbown/DeepSeek-TUI

RUST · +20,835 ★ / WEEK · TOTAL 26,400★

A terminal-based coding agent running on DeepSeek models. Part of the Claude Code alternative wave.

3

addyosmani/agent-skills

SHELL · +11,791 ★ / WEEK · TOTAL 40,400★

“Production-grade engineering skills for AI coding agents.” A collection of skills agents should ship with.

4

ruvnet/ruflo

TYPESCRIPT · +7,088 ★ / WEEK · TOTAL 49,700★

A Claude multi-agent orchestration platform. 49K★ accumulated since last year, +7K this week alone.

5

VectifyAI/PageIndex

PYTHON · +4,351 ★ / WEEK · TOTAL 30,800★

Reasoning-based RAG with no vectors. Index documents without standing up an embedding DB.

Four of the Top 5 (#1, #2, #3, #4) are all answers to the same question: how do you build AI agents, and where do you run them? A standard guide (Anthropic), local LLM (DeepSeek-TUI), shared skill set (agent-skills), orchestration (ruflo) — the whole week summarized in one line.

03 · HNHacker News this week — “local AI” and “layoffs”

HN #1
2,149↑
Hardware attestation is a monopoly weapon (GrapheneOS)
HN #2
1,840↑
Local AI needs to be the norm
HN #3
1,741↑
Valve releases Steam Controller CAD under CC
HN #4
1,350↑
Cloudflare cuts 20% of staff

HN ran in the same direction as GitHub. #2 “Local AI needs to be the norm” pulled 735 comments — the most-discussed thread of the week. How do we make locally-run AI the default? was the central question. Meanwhile Cloudflare’s 20% layoff (around 2,400 people) drew 988 comments — read as a signal that even AI-infra companies are shedding headcount right now.

04 · INCIDENTThis week’s security event — TanStack NPM supply-chain compromise

NPM SUPPLY-CHAIN HN 1,058↑ · 444💬

A TanStack maintainer account was hijacked and malicious packages were published to npm.

Cause: Suspected 2FA bypass + token leak. Blast radius: Exact download counts not disclosed. Lesson: The postmortem’s conclusion is that only projects already shipping supply-chain protections like npm provenance and OIDC token issuance could block and roll back in time.

TanStack (the team behind react-query and router) posted a postmortem on their official blog, and it landed in HN’s top 10. A package being safe yesterday doesn’t mean it’s safe today — the week was a reminder of exactly that.

This week was about “where and how do you run an agent.” Anthropic dropped its own standard guide, HN chanted local AI, and four of GitHub’s Top 5 were agent infrastructure.

— Eddie’s take

What to watch next week: whether anthropics/financial-services holds its trending spot, and how the DeepSeek-TUI vs. Claude Code local-agent race plays out.

Star counts and scores reflect the time of writing (2026-05-13 KST) and may shift within a week.


This post is part of an English mirror of a Korean dev experiments blog. Cross-posted for the global developer audience.




Leave a Comment