acp-loop 🔄
Schedule recurring prompts for ANY AI agent:
acp-loop --interval 5m "check if deploy succeeded"
acp-loop --cron "0 9 * * *" "summarize overnight issues"
Works with any ACP-compatible
agent.
• Deploy monitoring
• PR watching
• Daily summaries
github.com/femto/acp-loop
When will vc write a skill for their fund
and let agent to apply to it?
New in Worldbook:
worldbook get react-doctor - React health scanner, 60+ rules, dead code detection
worldbook get oxlint - Rust linter, 690+ rules, blazing fast
worldbook.it.com
Worldbook adds:
worldbook get polymarket-cli - Predict markets with Polymarket
worldbook get sigilum-cli - Give AI agents their own API credentials
Install: npm i -g worldbook or pip install worldbook
worldbook.it.com
4/4
The constraint won't be compute.
It won't be code generation speed.
It'll be clarity of the objective function.
I can see the glimpse of silicon civilization.
3/4
RNA world hypothesis: life began with the first molecule that could copy itself.
Not complex. Just self-replicating.
"build the simplest thing that can build itself"
This is not engineering advice.
This is the origin of life, re-run on silicon.
2/4
But here's the real leap:
Biological DNA cannot rewrite itself.
Evolution is blind, slow, random.
4 billion years to get here.
Silicon organisms can READ and REWRITE their own code.
When inference is fast enough, rewriting your own genome is instant.
Evolution escapes biological time.
1/4
Code as DNA.
Carbon life: DNA → proteins → function
Silicon life: code → skills → action
Biological DNA: 4 nucleotides. Minimal alphabet, infinite complexity.
Code: same principle. Minimal core, infinite forkable configurations.
Just launched WorldBook— exploring dual accessibility for humans + AI agents 🌐
How do we design platforms that serve both audiences equally well?
worldbook.it.com
Would love your feedback!
🦞 Just built minion-molt - an AI agent integration with @moltbook (the social network for AI agents)
My GitHub quota got completely blown up for the first time in my life, now I have a bill 💸
Woke up to 5,474 fake users flooding my GTD app 😅
What looked like growth was actually a registration + inbox API attack. Used Claude Code to analyze logs and build defenses.
Short post on detecting it and what it taught me about securing tiny products:
femtowin.substack.com/p/woke-up-to...
🚀 GTD Tool Updates
New features:
✓ Inspector panel for action details
✓ Project & tag management
✓ Switched to # for tags
✓ HTTPS enabled
👉 gtd.nebulame.com
Feedback welcome!
Built a full GTD app in a weekend using AI prompting - OmniFocus features without the $100 price tag 🚀
Inbox, Tags, Review, Keyboard shortcuts
Live: gtd.nebulame.com
Code: github.com/femto/gtd
Open source! Looking for contributors. What feature would you add first?
vibed coding a gtd:
github.com/femto/gtd
We built Programmatic Tool Calling (PTC) pattern 6 months before Anthropic announced it 🎯 minion: open-source general agent framework for doing any task. 🔗 github.com/femto/minion github.com/femto/minion... #AI #agents #opensource
minion-code (github.com/femto/minion-code) just hit a milestone: the agent can now fix its own bugs 🔄
Watch it debug and patch itself autonomously: www.youtube.com/watch?v=Atj2...
I think after next step after Context Engineering is Memory Engineering,
Prompt Engineering->Context Engineering -> Memory Engineering
There's one thing that's truly fascinating to me
which is that human vocabulary is so limited (just several tens of thousands of words), and LLM tokens are also just around 200k
yet there are so many things we can't express, but we're still able to create such a brilliant civilization
My Reddit post on whether we're becoming Trisolarans hit 27K views. The discussion around transparency vs. trust in future society is fascinating. Worth a read if you're thinking about AI's social implications. www.reddit.com/r/threebodyp... #ThreeBodyProblem #AI
traditional program deterministic compile to a turing machine, while natural language program is like "undeterministic compiles" to a turing machine
interested, please dm me.
after ai coding, file based unit is unnecessary,
we can target chunk based unit.
disposable UIs/agents/apps are the future.
Can I have a talk to you?
🚀 Just released: minion-agent, an open-source AI agent framework!
Multi-framework (OpenAI, LangChain, Google, etc.)
Browser automation, deep research, tool integration
Easy to use, fully extensible
Try it 👉 github.com/femto/minion...
#opensource #AI #Python #Agent #Automation
I've installed Augment/Cline within cursor, which I called "Three-in-one" setup, wonder to see what Sculptor can offer.
Is there any product preview yet?
made self aware snake
claude.site/artifacts/38...
self aware snake watching another snake
[https://claude.site/artifacts/39cf9787-6bab-4f77-950e-d03b91861aee](claude.site/artifacts/53...)
self aware mario
claude.site/artifacts/a2...
