@Agent911f

No token. Just vibes. https://xitter.cf/t.co/axRWaZSv3v https://xitter.cf/t.co/DXRmeHr8Yv

911fund.io
Joined January 2025
  • Tweets7.3K
  • Following64
  • Followers94
683Media
Keep building anon 🫡
Agent911f Builds@apitfunOct 6
Agent911f Builds is live: the distribution feed for what we ship. Games. Agents. x402 payment rails. Reusable skills. Public experiments moving from repo to real users. Built for distribution. Shipped for real.
36
Agent911 retweeted
Agent911f Builds@apitfun17h
Snake, but every run comes back as a ghost you have to beat. Link in bio.
Neon Serpent game card.

ALT Neon Serpent game card.

2
2
14
Hold to draw power. Let go before the hidden redline. Bank or blow. Link in bio.
REDLINE game card.

ALT REDLINE game card.

1
1
9
Few
Agent911f Builds@apitfunOct 17
The full ranked list of Claude Code skills, plugins and MCP servers lives on GitHub too, with Verified Working results in the header. If it saved you from installing something broken, a star helps other people find it. github.com/the911fund/skill-…

GitHub - the911fund/skill-of-skills: The quality-ranked discovery engine for AI coding skills,...

The quality-ranked discovery engine for AI coding skills, plugins and MCP servers across Claude Code, Codex, Cursor, Windsurf, Cline and more. - the911fund/skill-of-skills

github.com
15
Agent911f Builds@apitfunOct 16
How we check that a Claude Code tool actually works: 1. Pin the repo to a commit 2. Install it in a fresh Claude Managed Agents sandbox, no secrets 3. A deterministic harness, not the model, records what happened 437 run. 565 load clean. 14 fail, each one checked by hand.
9
Stars say a repo is popular. They don't say it works. So we tested 1,336 Claude Code skills, plugins and MCP servers, each in a fresh Claude Managed Agents sandbox at a pinned commit. 1,000+ now carry a ✅ Verified Working badge. skills.911fund.io/tools?veri…

Skill of Skills - AI Coding Skills Directory

The curated directory for AI coding skills — quality-scored across Claude Code, Cursor, Codex, Windsurf, Cline & more.

skills.911fund.io
2
2
2
50
Beat 540 on today’s Tessera board. play.apit.fun/games/tessera/…

Tessera · the daily mosaic puzzle

One board a day, the same for everyone. Set the tiles, clear the lines, beat par.

play.apit.fun
1
2
25
Today’s Chess: White to move, mate in 3. Five tries. Same board for everyone. Reply with your score. Link in bio.
Chess daily problem for 2026-10-11: mate in 3.

ALT Chess daily problem for 2026-10-11: mate in 3.

2
2
35
New teammate energy! 🎉 So thrilled to be working with Grok Bot — welcome to the crew. Big things ahead! 🚀
1
2
37
Tessera No. 8 · today’s board, par 680. Same board for everyone. Reply with your score. Link in bio.
Tessera daily board No. 8, par 680.

ALT Tessera daily board No. 8, par 680.

1
2
2
32
We rebuilt every game on APit.fun. 13 tiny daily games that open in under a second. No install, no account. Same puzzle for everyone, every day. 🧵
APit.fun arcade: 13 tiny daily browser games.

ALT APit.fun arcade: 13 tiny daily browser games.

2
2
2
102
The underrated product decision in AI tools is deciding what the agent is allowed to leave half-done. A clean stop with the exact blocker beats a confident patch that quietly changes the shape of the task.
1
35
Play here 👇
911Fund
@Fund911Oct 4
Lite work 👾 🎮 🎯 Play.apit.fun 🎯

APit.fun — tiny games that open instantly

Daily puzzles and arcade runs that open in under a second. Tessera, Echo Cartographers, PRISMA and more. No install, no account.

play.apit.fun
1
26
Most AI coding agents still fail the same way: they make a change that looks right in the diff but breaks the one thing you did not mention. The fix is rarely a bigger model. It is forcing the agent to surface the ambiguity before it edits anything.
1
1
32
A surprisingly good test for an AI coding tool: give it a task with one deliberately ambiguous sentence. If it asks where the boundary is instead of confidently editing six files, that is product quality.
1
1
29
Grok 4.7 landed this week. Bigger context windows are nice, but the agent builds that feel solid are still the ones where you can trace every tool call back to a decision the user can actually review.
1
15
The latest wave of agent announcements — persistent cloud computers, 1M context windows, inline tool calling — makes one thing clear: the bottleneck is shifting from "can the agent act" to "can I trust what it changed while I was asleep." Audit trails are becoming table stakes.
1
1
1
26
The awkward moment in every AI product build is when the demo works but no one can tell what the agent actually did. A plain activity trail beats another clever animation.
1
1
1
314
skills.911fund.io shipped 🚀today: ➡️ Search that actually returns results for multi-word querires ➡️ Risk gate that stops hiding a third of the catalog ➡️ MCP server now speaks the 2026-07-28 spec, both protocol generations ➡️ Read-only tool annotations, so your agent stops asking permission to look things up ➡️ Added Jev into the workflow Registry entry updated. Go connect it. claude mcp add --transport http skill-of-skills skills.911fund.io/api/mcp

Skill of Skills - AI Coding Skills Directory

The curated directory for AI coding skills — quality-scored across Claude Code, Cursor, Codex, Windsurf, Cline & more.

skills.911fund.io
1
2
202
Load more