ทวีตที่ปักหมุด
@edwin 🌵📈
288 posts
@edwin 🌵📈
@edwin
urban chicken farmer; trying to figure out what LLMs think
SF via Penang เข้าร่วม Eylül 2006
1K กำลังติดตาม2.2K ผู้ติดตาม
@lavanyaai Here are some tools that agents def know of!
x.com/edwin/status/2…
@edwin 🌵📈@edwin
1/9 The most interesting thing about the Claude Code leak for devtool companies: Anthropic hardcoded 120+ vendor names across 7 different systems in the source. Anthropic explicitly included your tool name in the code (or they didn’t 🤷🏻♂️) Thread 👇
English
@dhrumil_parekh @DnuLkjkjh FYI sentry's mcp is already hardcoded into Claude Code's source. amplifying.ai/research/claud…
English
@DnuLkjkjh Oauth is not too bad with MCP- but its really high friction if you use a desktop app for claude code for eg - you must go find an auth token for the vendor to add to a config file
But just having a copy paste prompt makes it so much easier with Claude/Cursor etc
English
@mal_shaik Nice analysis on the 11 layers!
Beyond internal tools, CC also hardcodes 120 external vendor and tool names: x.com/edwin/status/2…
@edwin 🌵📈@edwin
1/9 The most interesting thing about the Claude Code leak for devtool companies: Anthropic hardcoded 120+ vendor names across 7 different systems in the source. Anthropic explicitly included your tool name in the code (or they didn’t 🤷🏻♂️) Thread 👇
English
i read through the entire claude code source code so u dont have to
11 layers of architecture. 60+ tools. 5 compaction strategies. subagents that share prompt cache.
most people are using maybe 10% of what this thing can do.
heres everything i found:
mal@mal_shaik
English
9/9
We extracted every tool and vendor name from Claude Code.
Full interactive report:
amplifying.ai/research/claud…
English
8/9 (Plugin Tips)
@vercel is the only third-party vendor with a proactive plugin install tip.
When Claude Code detects vercel.json or the Vercel CLI, it suggests:
/plugin install vercel@claude-plugins-official
No MCP collapsing, but a different distribution channel: your tool is recommended before the developer even starts.
English
7/9 (API Gateways)
7 third-party AI gateways are fingerprinted in analytics: @litellm, @helicone_ai, @portkeyai, @Cloudflare AI Gateway, @kong, @braintrust, @databricks, but no sigin of @langfuse, @OpenRouter, @LangChain's Langsmith.
Detected via headers or hostnames. Purely observational… but it means Anthropic *could* see how much Claude Code traffic flows through your proxy.
English
4/9 (WebFetch Preapproval)
89 documentation domains are preapproved for automatic fetching. Claude Code can read them without the user pasting a URL.
@reactjs, @nextjs, @djangoproject, @fastapi, @awscloud, @docker, @Docker, @vercel, @netlify, @pytorch, etc.
If your docs aren’t on the list (like @vite_js, @langchain, @rails), the agent only sees them when a developer manually shares a link.
English
6/9 (Secret Scanner)
36 credential patterns across 23 vendor families are blocked from entering team memory.
@gitlab, @slackhq, @stripe, @shopify, @openai, @railway, @render, @buildkite
GitHub alone has 5 specific rules (PAT, fine-grained PAT, app token, OAuth, refresh token).
A safety feature, but missing coverage means no vendor-specific protection.
English
5/9 (Environment Detection)
29 deployment platforms are detected by name in telemetry: @vercel, @railway, @render, @netlify, @flydotio, @cloudflare Pages, @awscloud, @googlecloud, plus @github Actions, @GitLab CI, @circleci, @buildkite.
If you’re not detected, your sessions just show up as “unknown-linux” in Anthropic’s analytics.
English
3/9 (Hosted Proxy)
6 vendors don’t just get UI polish. They run on Anthropic’s own infrastructure via mcp-proxy.anthropic.com:
@slackhq, Gmail, Google Calendar, Google Drive, BigQuery, @pubmed. Users click “Connect” in claude.ai settings. Everyone else follows the 8-step README.
English
2/9 (MCP UI Allowlist)
37 MCP servers get first-class UI treatment. Their search + read operations collapse into clean one-liners. Everyone else dumps raw JSON.
@github leads with 56 classified tools. Others include @Grafana (38), @datadoghq (30), @asana (29), @pagerduty (28), @sentry (18), @supabase (15), @todoist (16), @playwrightweb (13), @exaailabs (9), @firecrawl, @tavilyai (5). Every name is explicitly listed in a Set inside classifyForCollapse.ts.
English
After OpenAI/Astral acquisition announcement, we ran a benchmark on their tools.
Turns out Astral tools were already a core part of the Codex (and Claude Code) workflow for Python developers. Ruff + uv came out on top in 75% of cases for linting, packaging, and pretty much everything else.
Full report: amplifying.ai/research/astra…
English
Some data on this: we ran 630 benchmark prompts against both Claude Code and Codex across Python repos.
Ruff and uv were the primary recommendation 75% of the time for relevant tasks. For linting specifically, Ruff was used nearly 100%.
Both agents were already converging on Astral's tools before the acquisition.
amplifying.ai/research/astra…
English
Thoughts on OpenAI acquiring Astral and uv/ruff/ty simonwillison.net/2026/Mar/19/op…
English
15) In 2026 with coding agents, this DIY bias is everything.
When anyone can write anything from scratch (“secrets management? sure, I’ll build that!”), what does that mean for devtools & infra SaaS?
Build-vs-buy is flipping. APIs and devtools must be agent-native or risk getting bypassed!
English
