Matt Dailey
@reactiverobot
Building http://ref.tools, previously @figma & @PalantirTech
قد يعجبك
100% agree, the strat they took to make it easy to develop servers worked great to get mcp off the ground but so much of the possibility is lost by clients being so tough pretty sure the team knows this tho
here's my gripes about mcp as an implementer - end users expect miracles from mcp and load in a bunch of them - most mcp servers are garbage, crash a lot, and the llm never bothers calling them - this whole "try sse and if it fails try streaming" is exactly why im so anti…
psst marketers, rather that writing a "us vs them" page you should write a Claude Skill page that just helps me migrate from them to you
Ref October highlights: - 𝗿𝗲𝗮𝗱_𝘂𝗿𝗹 has more cache hits and is much more stable scraping pages on the fly. - Started SOC2 with Vanta - New documentation site docs.ref.tools Also MCP Developers Summit! All the talks are up youtube.com/@MCPDevSummit
who is working on prompt injection detection for coding agents? what should I read? who should i follow? i notice some agents are way better than others. For example, claude code will often ignore subtle manipulation and decline overt attempts but some other agents wont
Stoked to demo ref.tools on the closing night of GitHub Universe! Come hangout and see cool stuff When: October 29th Where: San Francisco Brewing Co. Demos from @floxdevelopment, @AI21Labs, @github, @boundaryML and yours truly! luma.com/5lwci3p1
Wrapped up @ExaAILabs and ref.tools into a @FactoryAI custom command Now the Droids have a custom command to perform planning research and find reference links Output is packed in a result envelope for the Droid to render results, cache, and chain follow-up steps
can some one link a good explanation of passkeys? i feel like they just started showing up every where and i just clickyclickclick and its setup and im authenticated but how is that more secure? or is it even?
See ya at GitHub Universe!
hey that's me!
My talk from @mcpsummit is live! Check it out: youtube.com/watch?v=cGki4V…
youtube.com
YouTube
[Session] Improving Agentic Search with MCP Sessions with Matt Dailey...
background agents have warped my brain i can ONLY multi-task now
🚀 Launch week feature 2 Agentic search in the web app Ref is primarily an MCP server but you can also use ref.tools/search This is one of the ways new users test Ref BUT it used work differently and be kind of bad. Now it works just like the MCP server and is good
🚀 Launch week feature 1 Upload markdown files to your index I recently learned some enterprising users were converting markdown files to PDF and uploading to Ref to take advantage of the token-efficient search. Yall don't need to do that. Now you can just upload the .md 😅
🚀 LAUNCH WEEK DAY 3! This is the last before @mcpsummit 🚀 👯 Teams You can now create a shared index of context for your team with RBAC and unified billing. Plus you can migrate can easily migrate your existing personal index to get started. try at ref.tools/create-team
do scheduled posts on x just not work? i scheduled all my launch week posts and they all just disappeared when the time came 😅
United States الاتجاهات
- 1. Veterans Day 387K posts
- 2. Woody 13.7K posts
- 3. Tangle and Whisper 2,108 posts
- 4. #stateofplay 11K posts
- 5. Toy Story 5 19K posts
- 6. Nico 144K posts
- 7. Luka 85.3K posts
- 8. Gambit 43.9K posts
- 9. Travis Hunter 4,056 posts
- 10. Payne 12.5K posts
- 11. Tish 5,713 posts
- 12. SBMM 1,588 posts
- 13. Square Enix 5,453 posts
- 14. De Minaur 3,261 posts
- 15. Wike 127K posts
- 16. Sabonis 3,939 posts
- 17. Mavs 33.3K posts
- 18. Antifa 194K posts
- 19. Vets 34.6K posts
- 20. Tokon 1,813 posts
Something went wrong.
Something went wrong.