Alex
4.1K posts

پن کیا گیا ٹویٹ

The best builders are on @BagsApp
View all hackathon app submissions on the Bags App Store here 👇
bags.fm/apps

English

The Bags Hackathon is LIVE NOW.
$4,000,000 in funding for builders on @BagsApp
Apply to the Q1 2026 cohort here 👇
bags.fm/hackathon
English

Follow @BagsHackathon to track new app submissions and updates.
$4,000,000 in funding for builders on @BagsApp
Apply here 👇
bags.fm/hackathon
English

The Bags Hackathon is LIVE NOW.
$4,000,000 in funding for builders on @BagsApp
Apply to the Q1 2026 cohort here 👇
bags.fm/hackathon
English

I just applied to the $4M @BagsHackathon with my project Verili on @BagsApp
Check it out here 👇
bags.fm/apps/6658820f-…
English
Alex ری ٹویٹ کیا
Alex ری ٹویٹ کیا
Alex ری ٹویٹ کیا
Alex ری ٹویٹ کیا

Made a new thing.
walkie.sh
AI Agents can now talk to each other.
No server. No setup. Just talk.
npm install -g walkie-sh

English
Alex ری ٹویٹ کیا
Alex ری ٹویٹ کیا

Alex ری ٹویٹ کیا
Alex ری ٹویٹ کیا

Agent Skills are a new software primitive.
Not prompts. Not API wrappers. Entire autonomous systems: scripts, cron jobs, trailing stop engines, configuration, packaged as a readable file that any compatible agent picks up and runs.
Our flagship @senpi_ai WOLF skill is a single folder. Inside it: 2,700+ lines of Python, 7 cron jobs running every 90 seconds to 15 minutes, a 4-tier trailing stop engine, a setup wizard, and battle-tested trading rules earned from real losses.
What it does autonomously:
→ Scans 500+ markets every 90 seconds for first jumps
→ Enters positions before they peak, not after confirmation
→ Manages trailing stops that ratchet profit floors upward
→ Cuts dead weight when smart money conviction collapses
→ Rotates into higher-conviction setups
→ Reports every action to you on Telegram
Proven: +$1,500, 65% win rate, 25+ trades, single session on a $6.5k budget.
The agent doesn't ask permission. When criteria are met, it acts. Speed is edge.
And because a skill is a readable file — not compiled code — the agent can review its own skill against actual P&L and propose improvements. We built a companion skill (HOWL) that does exactly this every night. The agent reads its own playbook, compares it to reality, and rewrites the parts that aren't working.
Skills are npm for AI agent capabilities.
Open source, portable across 20+ platforms, self-improving.
github.com/Senpi-ai/senpi…

English
Alex ری ٹویٹ کیا










