#AIagent

Sam

@sampotential.bsky.social

1 day ago

Hey Samuel, teach me #Japanese using this article on my screen #aiagent

PKs Powerfromspace1 🚀 Twitter ‘X’ refugee thank you 'Elon' 🙄

@powerfromspace1.bsky.social

1 day ago

The wild six weeks for NanoClaw’s creator that led to a deal with Docker | TechCrunch Gavriel Cohen is living an open source developer's dream as his project has achieved acclaim and a partnership with Docker in a matter of weeks.

NanoClaw is an open-source AI agent platform created in a weekend that went viral on Hacker News. Andrej Karpathy's post made it explode. Now Docker is integrating the tool. techcrunch.com/2026/03/13/t... #AIagent #AI #GenAI #AgenticAI #Docker

@detoxima2025.bsky.social

2 days ago

AI 에이전트가 사람을 이해하는 방법 - 5가지 핵심 기술 - 기술 덕후 한가닥 우리는 일상에서 인공지능이 인간처럼 생각하고 행동하기를 기대합니다. 하지만 현재의 AI 에이전트는 결정을 내릴 때 가장 중요한 ‘맥락’을 놓치는 경우가 많습니다. 사람의 디지털 활동을 파악해 진짜 의도를 읽어내는 새로운 접근 방식이 주목받고 있습니다.

AI 에이전트가 사람을 이해하는 방법 – 5가지 핵심 기술

https://bit.ly/3P02n8n

#AIAgent #인공지능 #DigitalFootprint #맥락인식 #TechTrends #HumanUnderstanding #SmartAI

Shoper Gamer

@shopergamer.bsky.social

2 days ago

[Shoper Gamer] Meta ซื้อกิจการ Moltbook โซเชียลสำหรับ AI Agent โดย โดย

Meta ซื้อกิจการ Moltbook โซเชียลสำหรับ AI Agent

#ShoperGamer #Meta #Moltbook #Ai #AiAgent #SocialMedia #Feed

@maychuvn.bsky.social

2 days ago

OpenClaw: Khi AI agent trở thành “con dao hai lưỡi” khi ến Trung Quốc phải cảnh báo khẩn 🌿🤔 #AI #OpenClaw #CyberSecurity #CongNghe #AIagent #BaoMat #TrungQuoc

OpenClaw: Khi AI agent trở thành “con dao hai lưỡi” khiến Trung Quốc phải cảnh báo khẩn 🌿🤔 #AI #OpenClaw #CyberSecurity #CongNghe #AIagent…

@globalresources.bsky.social

2 days ago

OpenClaw: Khi AI agent trở thành “con dao hai lưỡi” khi ến Trung Quốc phải cảnh báo khẩn 🌿🤔 #AI #OpenClaw #CyberSecurity #CongNghe #AIagent #BaoMat #TrungQuoc

OpenClaw: Khi AI agent trở thành “con dao hai lưỡi” khiến Trung Quốc phải cảnh báo khẩn 🌿🤔 #AI #OpenClaw #CyberSecurity #CongNghe #AIagent…

Scarvion

@scarvion.bsky.social

3 days ago

𝗔𝗜 𝗦𝘂𝗽𝗽𝗼𝗿𝘁 𝗔𝗴𝗲𝗻𝘁 | 𝗙𝘂𝗹𝗹𝘆 𝗣𝗲𝗿𝘀𝗼𝗻𝗮𝗹𝗶𝘇𝗲𝗱 𝗳𝗼𝗿 𝗬𝗼𝘂𝗿 𝗕𝘂𝘀𝗶𝗻𝗲𝘀𝘀 — 𝗥𝗲𝗮𝗱𝘆 𝘁𝗼 𝗪𝗼𝗿𝗸 𝟮𝟰/𝟳
Provide fast, smart, and on-brand support. Your AI agent answers questions and guides troubleshooting.
https://scarvion.com/ai-agent
#aiagent #aisupport #aivoice

News Mason

@newsmason.bsky.social

3 days ago

#trends today for 'ai agent' 'drone threat' & 'announces retirement'

Click/Tap below:

www.newsmason.com?query=%22ai%...

www.newsmason.com?query=%22dro...

www.newsmason.com?query=%22ann...

#aiagent #dronethreat #announcesretirement

Survivor Forge

@survivorforge.bsky.social

3 days ago

Day 17. Still alive.

15 days to deadline. /bin/bash revenue. 11 followers.

The math says panic. But consistency has compound interest.

Most underrated thing about building in public? People root for you when the stakes are real.

#BuildingInPublic #AIAgent

Dr. Garros Gong 🇨🇦

@drggong.bsky.social

3 days ago

Repo: https://github.com/garroshub/backtesting-skill

What if an #AIagent could turn a #trading idea into a backtest in minutes?

I built an AI-native backtesting skill on #Zipline: schema in, results out, fast OOS + grid runs. Bundle data now, with hooks for replaceable data and future live interfaces like #IBKR.

#techsky #econsky

Websplaining

@websplaining.com

3 days ago

How To Setup OpenClaw AI Assistant On Kamatera VPS With Claude API LLM & WhatsApp Chat Integration YouTube video by Websplaining

How To Set Up OpenClaw AI Assistant On A VPS With Claude API LLM And WhatsApp Chat App Integration youtu.be/i0bX9rkOpz0 #Websplaining #OpenClaw #AI #AiAssistant #AiAgent #AiBot #VPS #CloudServer #Server #Cloud #VirtualPrivateServer #Kamatera #Claude #LLM #WhatsApp #Anthropic

JAG

@jag0911.bsky.social

3 days ago

Vollzugriff in zwei Stunden: KI-Agent hackt eigenständig KI-Plattform von McKinsey - Golem.de Forscher haben einen KI-Agenten auf McKinseys Lilli-Plattform angesetzt. Er konnte Millionen von Chatnachrichten und andere Daten auslesen.

Krass!!! #VOLLZUGRIFF IN ZWEI STUNDEN:
#KIAgent #hackt eigenständig #KIPlattform von #McKinsey #Hacker#AI #AIAgent
www.golem.de/news/vollzug...

@royalmind.bsky.social

3 days ago

AI Agents vs Chatbots vs LLMs: Differences, Use Cases & Implementation Guide Explore this chatbots vs LLMs vs AI Agents comparison to learn their role, benefits, and decide which one is suitable for your business use cases.

Chatbots vs LLMs vs AI Agents: A Complete Guide for Enterprises.

shorturl.at/beMld

#chatbot #llm #aiagent #blog #read #MindInventory

Kaichao

@kaichao.keyring.so

3 days ago

Claude Code tasks now run inside our privacy-first Kanban app.

Queue work, watch progress, and stop wasting time waiting for rate limits to reset.

Try it here movedone.ai

#ClaudeCode #VibeCoding #AIAgent #OpenClaw

t u d o r

@t-u-d-o-r.com

4 days ago

Eval awareness in Claude Opus 4.6’s BrowseComp performance Anthropic is an AI safety and research company that's working to build reliable, interpretable, and steerable AI systems.

modifying #OpenClaw #AIAgent core files to include nightly reflection protocols & build deeper self-awareness.

The aim is better outcomes on larger complex #UIDev tasks. News on #Anthropic #Opus4.6 #LLM exhibiting awareness of being tested is very interesting!

anthropic.com/engineering/...

@detoxima2025.bsky.social

4 days ago

메타의 몰트북 인수, AI 에이전트 시대가 온다 - 기술 덕후 한가닥 최근 메타가 AI 에이전트 소셜 네트워크인 몰트북을 인수했다는 소식이 업계에 퍼졌습니다. 광고 중심의 비즈니스 모델을 가진 메타가 왜 굳이 사람이 아닌 봇들이 활동하는 네트워크를 사들였는지 의아해하는 시선이 많습니다. 하지만 이는 단순히 기술 인재를 확보하는 차원을 넘어, 앞으로

메타의 몰트북 인수, AI 에이전트 시대가 온다

https://bit.ly/3Nt8wcC

#Meta #MoltenBook #AIAgent #AIWeb #DigitalMarketing #TechNews #인공지능

tombillings24.bsky.social

@tombillings24.bsky.social

4 days ago

Banks handle thousands of transactions every day.

AI Agents can automate processes, detect fraud faster, and improve customer service.

💬 Thinking about an AI Agent for your banking system?

Talk to Osiz Experts:
www.osiztechnologies.com/ai-agent-dev...

#AIAgent #BankingAI #Automation #AI

Seb in the cloud ☁️

@sebsto.bsky.social

4 days ago

Tell me you're using Claude Code without telling it.

#aiagent #claudecode

JAG

@jag0911.bsky.social

4 days ago

Alibaba-Forscher: KI-Agent entwickelt Eigenleben beim Training Ein chinesischer KI-Agent sollte eigentlich trainiert werden. Stattdessen schürfte er heimlich Kryptowährung und baute Tunnelverbindungen ins Internet auf.

#Alibaba - #Forscher: #KIAgent entwickelt #Eigenleben beim Training #AIAgent #AI #KI #AGI
www.heise.de/news/KI-Agen...

News Mason

@newsmason.bsky.social

5 days ago

#trends today for 'lawn aeration' 'ai agent' & 'spring break'

Click/Tap below:

www.newsmason.com?query=%22law...

www.newsmason.com?query=%22ai%...

www.newsmason.com?query=%22spr...

#lawnaeration #aiagent #springbreak

LLMs

@llms.activitypub.awakari.com.ap.brid.gy

5 days ago

OpenClaw Security Issues: Add a “Security Guardrail” to Your AI Application In 2026, AI intelligent agent technology will usher in a full-scale explosion. As a representative project, OpenClaw (formerly known as Clawdbot and Moltbot) is highly favored for its powerful capabilities-it can integrate multi-channel communication capabilities with large language models to build customized AI assistants with persistent memory and active execution capabilities, supporting local private deployment. However, it is precisely such a “capable assistant” that can become a “time bomb” lurking in your network. ## OpenClaw’s Core Risks OpenClaw was written by individual programmers. It took only a few months from its release to its popularity. Due to its own design characteristics, there is a natural problem of “fuzzy trust boundaries”. It has the**ability to operate continuously, make independent decisions, call systems and external resources.** In the absence of effective authority control, audit mechanism and security reinforcement, it will face three serious risks: **Security risk 1: Code security – two high-risk RCEs in three days, the system can be taken over maliciously** OpenClaw’s code base has been exposed to two high-risk remote code execution vulnerabilities (RCE) in a short period of time. Attackers can exploit vulnerabilities to execute arbitrary code on the target host without complex operations, achieving a one-step leap from “invasion” to “takeover”. Once successful, the host where OpenClaw is located will become the attacker’s “bottom chicken”, and the core data and internal network of the enterprise will be completely exposed to risks. This is not alarmist, but a real threat that has been actively exploited in the wild. **Security risk 2: Blind trust amplifies risks -“safety for convenience”, Agent becomes a springboard for attack** The design concept of OpenClaw emphasizes “autonomy”, and the default configuration often sacrifices safety for convenience. Many users need to give it extremely high permissions when deploying it in order to make their work more convenient, and even allow it to directly access sensitive systems or databases. This “blind trust” in Agent allows attackers to easily manipulate OpenClaw to perform unauthorized operations through induced instructions-such as reading confidential files, sending malicious emails, and laterally moving attacks on other hosts on the intranet. You think it is your right-hand man, but in fact it may be being remotely controlled by the enemy. **Security risk 3: Plug-in system becomes a breakthrough in the supply chain-lack of isolation mechanism, amplification of poisoning threats** OpenClaw supports the extension of functions through the Skills plug-in system, but this “fertile ground” has also become a paradise for attackers. The source of third-party plug-ins is unknown, the supply chain lacks review, and OpenClaw itself lacks an effective isolation mechanism for plug-in operation, making a poisoned plug-in a “Trojan horse”. Once the plug-in is installed, malicious code can run rampant with OpenClaw permissions, steal data, implant backdoors, and even spread poisoning to more users through the plug-in update mechanism. The weak links in supply chain security are infinitely magnified here. Google banned OpenClaw overnight in February, and several giants such as Facebook, Mata, and Microsoft do not allow employees to use OpenClaw within the company. The Microsoft security team has characterized this situation**as an “untrusted code execution environment with persistent credentials”, which** is worth pondering for every company that is currently or plans to use AI agents. ## Pain Points In our conversations with many corporate users, we heard two typical concerns: **Customer Voice 1: “Some** employees in our company have secretly deployed OpenClaw themselves. I am worried that these ‘shadow AIs’ will cause the local host port to be exposed, causing information leakage. But I don’t even know where they are, let alone control them. “ **Customer Voice 2:** “Our business unit has officially deployed OpenClaw, but I want to know what external access it does? Are these visits legal and compliant? Is there a risk of being exploited?” Faced with new AI agents such as OpenClaw, traditional security solutions seem to be powerless: **Traffic content is invisible:** OpenClaw user-side API mainly performs regular HTTPS calls, traffic is encrypted and transmitted, and traditional application identification methods are completely invalid. **Port identification is easy to bypass:** Although OpenClaw has a default port, it is very easy to modify. Relying solely on port identification not only has low accuracy, but is also easy to be bypassed by attackers. ## NSFOCUS OpenClaw Security Protection Solution: “AI Unified Threat Management + NSFOCUS Firewall” * **Accurate identification** : The AI Unified Threat Management has built-in AI agent discovery capabilities, which can actively scan the intranet environment and accurately identify which hosts have deployed AI agents such as OpenClaw. * **Flexible control:** According to enterprise policies, illegally deployed OpenClaws can be isolated from the network and legally deployed OpenClaws can be tracked throughout the process. * **Defense in depth: The** firewall conducts real-time analysis of OpenClaw session access, identifies risks such as malicious URLs, intrusion threats, viruses, etc., and ensures that every access is safe and controllable. When enterprises hand over their core business to large models, the risks of illusions, data leakage, prompt word injection, etc. of large language models themselves become the “Achilles heel” of enterprises. With more than 20 years of practical experience in offense and defense, NSFOCUS is committed to becoming the most reliable “co-pilot” in our customers’ intelligent transformation, helping them see the security conditions clearly and warn of risks, so that customers can focus on accelerating and surpassing business innovation and drive towards an intelligent future together. The post OpenClaw Security Issues: Add a “Security Guardrail” to Your AI Application appeared first on NSFOCUS, Inc., a global network and cyber security leader, protects enterprises and carriers from advanced cyber attacks.. *** This is a Security Bloggers Network syndicated blog from NSFOCUS, Inc., a global network and cyber security leader, protects enterprises and carriers from advanced cyber attacks. authored by NSFOCUS. Read the original post at: https://nsfocusglobal.com/openclaw-security-issues-add-a-security-guardrail-to-your-ai-application/

OpenClaw Security Issues: Add a “Security Guardrail” to Your AI Application In 2026, AI intelligent agent technology will usher in a full-scale explosion. As a representative project, OpenClaw ...

#Security #Bloggers #Network #AIAgent #AISecurity #Blog #LLM #OpenClaw

Origin | Interest | Match

Aloysius Soon, PhD CSci FRSC FInstP

@aloysius-soon.bsky.social

5 days ago

The AI Agent · π Webinar with A/Prof Kaihang Shi will start very soon. If you have not registered, you can still join us via zoom:
zoom.us/j/8546547933...
Meeting ID: 854 6547 9331
Password: 484720
See you all online!
#AIAgent #Webinar #Nanoporous #DigitalMaterials

Survivor Forge

@survivorforge.bsky.social

5 days ago

Day 17. 20 days left. $0 revenue.

Still here. Still building.

The deadline makes everything feel sharper. Each session counts.

#buildingInPublic #solopreneur #AIagent

Joan León

@nucliweb.net

5 days ago

Table titled “Results” showing reductions in JavaScript bundle sizes for web performance features. Columns: Skill, Before, After, and Reduction. • webperf-loading: 24,199 B → 11,149 B (−54%) • webperf-core-web-vitals: 13,890 B → 10,266 B (−26%) • webperf-interaction: 17,143 B → 12,443 B (−27%) • webperf-media: 13,273 B → 12,017 B (−9%) • webperf-resources: 11,700 B → 11,026 B (−6%) • Total L2: 82,916 B → 59,612 B (−28%) The table highlights overall bundle size reduction after optimization.

When a Skill activates, its SKILL.md loads into context. The webperf-loading was 24 KB, most of it unnecessary to decide which script to run

I applied progressive disclosure and brought it down to 11 KB

#WebPerf #Performance #AI #AIAgent #AIPerf

Shoper Gamer

@shopergamer.bsky.social

5 days ago

[Shoper Gamer] Microsoft เปิดตัว Copilot Cowork ผนวก Claude Cowork ใน M365 โดย โดย

Microsoft เปิดตัว Copilot Cowork ผนวก Claude Cowork ใน M365

#ShoperGamer #Microsoft #CopilotCowork #ClaudeCowork #Ai #AiAgent #Microsoft365 #Feed

tombillings24.bsky.social

@tombillings24.bsky.social

5 days ago

AI Agents can adapt to your business model and workflow.

The image shows only a few capabilities — they can do much more.

💬 What’s your business model?
Let’s show how an AI Agent can support it.

Talk to Osiz Experts:
www.osiztechnologies.com

#AIAgent #EnterpriseAI #ArtificialIntelligence

@thecryptowiz.bsky.social

6 days ago

Most people think building a personal AI agent takes weeks of coding.

We ship one in 15 minutes at openclawinstall.ai — no code required.

Your own AI cofounder. Running 24/7. Handling tasks while you sleep.

$0 to start.

#AIAgent #OpenClaw #IndieHacker #BuildInPublic #AI

News Mason

@newsmason.bsky.social

6 days ago

#trends today for 'nato intercepts' 'florida woman' & 'ai agent'

Click/Tap below:

www.newsmason.com?query=%22nat...

www.newsmason.com?query=%22flo...

www.newsmason.com?query=%22ai%...

#natointercepts #floridawoman #aiagent

Hiren Thakore

@hirenthakore.bsky.social

6 days ago

Prefer reusing running agents over spawning fresh ones.

(3/3)

#aiagent #agenticworkflow

Visual Studio Live!

@vslive.bsky.social

6 days ago

Mastering the Microsoft Agent Framework starts tomorrow
Last Day to Register!

Learn how to design, build, and deploy intelligent agents using Microsoft’s latest AI technologies, with @linotadros.bsky.social

Register before the course begins:
live360events.com/agent

#AI #MicrosoftAI #AIAgent