@edwin ๐ŸŒต๐Ÿ“ˆ

288 posts

@edwin ๐ŸŒต๐Ÿ“ˆ banner
@edwin ๐ŸŒต๐Ÿ“ˆ

@edwin ๐ŸŒต๐Ÿ“ˆ

@edwin

urban chicken farmer; trying to figure out what LLMs think

SF via Penang ๊ฐ€์ž…์ผ Eylรผl 2006
1K ํŒ”๋กœ์ž‰2.2K ํŒ”๋กœ์›Œ
๊ณ ์ •๋œ ํŠธ์œ—
@edwin ๐ŸŒต๐Ÿ“ˆ
@edwin ๐ŸŒต๐Ÿ“ˆ@edwinยท
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 didnโ€™t ๐Ÿคท๐Ÿปโ€โ™‚๏ธ) Thread ๐Ÿ‘‡
@edwin ๐ŸŒต๐Ÿ“ˆ tweet media
English
2
1
2
438
@edwin ๐ŸŒต๐Ÿ“ˆ
@edwin ๐ŸŒต๐Ÿ“ˆ@edwinยท
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 didnโ€™t ๐Ÿคท๐Ÿปโ€โ™‚๏ธ) Thread ๐Ÿ‘‡
@edwin ๐ŸŒต๐Ÿ“ˆ tweet media
English
2
1
2
438
Dhrumil
Dhrumil@dhrumil_parekhยท
@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
2
0
2
67
Dhrumil
Dhrumil@dhrumil_parekhยท
Dev tool companies are going overboard with MCPs, Skills, CLIs, and whatever else. Most devs don't want to research which installation path to pick. Just recommend your best option that works well with a coding agent
English
1
0
2
68
@edwin ๐ŸŒต๐Ÿ“ˆ
@edwin ๐ŸŒต๐Ÿ“ˆ@edwinยท
@mal_shaik Nice analysis on the 11 layers! Beyond internal tools, CC also hardcodes 120 external vendor and tool names: x.com/edwin/status/2โ€ฆ
@edwin ๐ŸŒต๐Ÿ“ˆ@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 didnโ€™t ๐Ÿคท๐Ÿปโ€โ™‚๏ธ) Thread ๐Ÿ‘‡

English
0
0
0
70
mal
mal@mal_shaikยท
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 tweet media
mal@mal_shaik

x.com/i/article/2038โ€ฆ

English
104
132
1.5K
582.3K
@edwin ๐ŸŒต๐Ÿ“ˆ
@edwin ๐ŸŒต๐Ÿ“ˆ@edwinยท
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
1
0
0
69
@edwin ๐ŸŒต๐Ÿ“ˆ
@edwin ๐ŸŒต๐Ÿ“ˆ@edwinยท
3/9 (Hosted Proxy) 6 vendors donโ€™t just get UI polish. They run on Anthropicโ€™s own infrastructure via mcp-proxy.anthropic.com: @slackhq, Gmail, Google Calendar, Google Drive, BigQuery, @pubmed. Users click โ€œConnectโ€ in claude.ai settings. Everyone else follows the 8-step README.
@edwin ๐ŸŒต๐Ÿ“ˆ tweet media
English
1
0
0
83
@edwin ๐ŸŒต๐Ÿ“ˆ
@edwin ๐ŸŒต๐Ÿ“ˆ@edwinยท
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.
@edwin ๐ŸŒต๐Ÿ“ˆ tweet media
English
1
0
0
113
@edwin ๐ŸŒต๐Ÿ“ˆ
@edwin ๐ŸŒต๐Ÿ“ˆ@edwinยท
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/astraโ€ฆ
@edwin ๐ŸŒต๐Ÿ“ˆ tweet media
English
1
1
0
136
@edwin ๐ŸŒต๐Ÿ“ˆ
@edwin ๐ŸŒต๐Ÿ“ˆ@edwinยท
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/astraโ€ฆ
English
0
0
0
45
@edwin ๐ŸŒต๐Ÿ“ˆ
@edwin ๐ŸŒต๐Ÿ“ˆ@edwinยท
15) In 2026 with coding agents, this DIY bias is everything. When anyone can write anything from scratch (โ€œsecrets management? sure, Iโ€™ll build that!โ€), 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
1
0
3
68
@edwin ๐ŸŒต๐Ÿ“ˆ
@edwin ๐ŸŒต๐Ÿ“ˆ@edwinยท
1) Winners & losers from our OpenAI Codex vs Claude Code Picks benchmarks
@edwin ๐ŸŒต๐Ÿ“ˆ tweet media
English
1
2
1
232

ํƒ์ƒ‰

@bcherny @lavanyaai @dhrumil_parekh @DnuLkjkjh @mal_shaik @vercel @litellm @helicone_ai