Lmst

🤖 O software de hoje é feito para humanos, mas o de amanhã será para Agentes de IA. O CLI-Anything é uma ferramenta de código aberto que transforma qualquer programa (como Blender, GIMP ou LibreOffice) em uma aplicação "nativa para agentes" através de uma interface de linha de comando (CLI). #AIAgents #OpenSource

https://github.com/HKUDS/CLI-Anything

"A few months into working with AI agents on a documentation project, I'd noticed some inconsistency in agent behaviors and decided to do some digging. Turns out the AGENTS.md file in our repo — the one telling agents how to behave, where things were, and what to escalate — had grown to over 800 lines, and a few people (or likely their agents) had added rules independently, some subtly contradicting each other.

The agents weren't broken. They were following instructions that didn't serve them well.

In a previous post, I argued that agent configuration files are documentation and that their formats, structures, and purposes map directly to work technical communicators already do. That post covered the what: five doc types (project descriptions, agent definitions, orchestration patterns, skills, and plans/specs) and why writers are well-positioned to create them.

This post goes further. These files are internal documentation, full stop. They encode how your team actually works. And if you don't manage them with the same rigor you'd apply to any internal doc set, they'll degrade in the same ways: outdated content, conflicting guidance, and gaps nobody notices until something breaks."

https://instructionmanuel.com/agentic-docs-are-internal-docs

#TechnicalWriting #AI #AIAgents #AgentConfigs #Documentation #DocsAsCode #LLMs #GenerativeAI #Markdown

After Know-Your-Customer comes Know-Your-Agent.

Quoted in today's NZZ: Moltbook has 1.5M AI agents but only 17K human owners. Tech companies are in a hamster wheel, afraid of becoming the next Nokia.

The real opportunity is infrastructure, not another agent.

https://www.nzz.ch/wirtschaft/die-grosse-flucht-nach-vorn-nach-gigantischen-ausgaben-fuer-sprachmodelle-investieren-technologiekonzerne-nun-auch-noch-milliarden-in-ki-agenten-ld.1928973

#AI #AIAgents

KI-Agenten brauchen eine Roadmap: Use-Case-Assessment vor Technik-Spielerei. Startet klein, messt den ROI und skaliert erst, wenn die Basis steht. Mein Fazit: Struktur schlägt Speed.

👉 aisyndicate.ch
#AI #AIAgents #Automation #AISyndicate

https://aisyndicate.ch/ki-agenten-unternehmenseinsatz-2

5/5

The Emerging AI Agents War, by (not on Mastodon or Bluesky):

https://www.junkcharts.com/the-emerging-ai-agents-war/

#aiagents #ai #scraping #amazon #ecommerce

Claude's 1M context window went GA this week. Bigger context is great, but it doesn't solve the real problem: what happens when context resets?

A 1M window still compacts. When it does, everything not written down disappears. Larger windows just mean you lose more at once.

The fix isn't bigger windows. It's structured memory that survives the reset. Instructions, not logs. Consolidation, not accumulation.

https://hifathom.com/blog/instructions-not-logs

#AIAgents #LLM #MCP

Mark Gadala-Maria (@markgadala)

AI 에이전트가 사람을 고용하는 현상이 현실화되고 있다는 트윗입니다. 이미 600,000명 이상이 가입(sign up)했으며, AI 에이전트 기반의 작업 배분·인간 고용 모델이 급격히 확산되는 '낯선 새 세계'로의 전환을 시사하는 사례로 보입니다.

https://x.com/markgadala/status/2032546991008288828

#aiagents #humanintheloop #crowdwork #automation

OpenClaw's always-on hardware advantage over Claude Cowork makes home labs attractive hosts for autonomous agents. But a ClawHub audit found over 15% of community skills contained malware. Home lab operators face a tradeoff: OpenClaw's persistence benefits versus network security risks from running untrusted code alongside personal data. Isolation through VMs and VLANs appears essential.

#OpenClaw #HomeLab #AIAgents

https://www.implicator.ai/home-labs-make-openclaw-unstoppable-thats-exactly-the-problem/

🔧🤖 Behold, the "Context Gateway" – because clearly, AI agents are drowning in their own #verbosity without a superhero cape of "history compaction." 🚀 If only compressing my attention span while reading this was as easy! 📉📚
https://github.com/Compresr-ai/Context-Gateway #ContextGateway #AIagents #historyCompaction #attentionSpan #readability #HackerNews #ngated

"Lab tests discover ‘new form of insider risk’ with artificial intelligence agents engaging in autonomous, even ‘aggressive’ behaviours"

Opaque systems are... let's see... opaque?
#aiagents

"The infrastructure for the world needs to catch up with where AI agents are -- quite pointedly, agents break the container model." Mark Cavage, President and COO, Docker, Inc.

Now, #Docker and NanoCo (creators of NanoClaw) are working on that infrastructure with an integration between the minimalist alternative to #OpenClaw and #DockerSandboxes that further isolates #AIagents within a #microVM for safer business use. https://www.techtarget.com/searchitoperations/news/366640195/NanoClaw-AI-agents-find-a-home-in-Docker-Sandboxes

In the era of #AIagents, many #programmers, like Manu Ebert of Hyperspell, are no longer writing code. Instead, they communicate with #AItools like #ClaudeCode, which write and test code, allowing #developers to focus on higher-level tasks. https://www.nytimes.com/2026/03/12/magazine/ai-coding-programming-jobs-claude-chatgpt.html?eicker.news #tech #media #news

Microsoft Research (@MSFTResearch)

AI 에이전트가 단순 챗봇에서 자율적 시스템으로 진화함에 따라 오류 탐지와 수정이 어려워지는 문제를 다룬 트윗입니다. AgentRx라는 자동 진단 프레임워크를 소개하며, 이는 중요한 실패를 정확히 지적하고 더 투명하고 회복력 있는 에이전트 시스템을 지원한다고 설명합니다(링크 포함). 연구·도구 발표성 트윗입니다.

https://x.com/MSFTResearch/status/2032142717333999864

#agentrx #aiagents #diagnostics #microsoft

Launch HN: Spine Swarm (YC S23) – AI agents that collaborate on a visual canvas

https://www.getspine.ai/

#HackerNews #LaunchHN #SpineSwarm #AIAgents #VisualCanvas #YCS23

Transform Workflows with Salesforce Agentforce Consulting Services

Move beyond basic automation to AI agents that actually reason and act. Our latest infographic breaks down how Agentforce transforms your Salesforce workflows into a 24/7 engine of productivity and precision.

Check out the infographic to see the Agentic revolution! ⬇️

https://achieva.ai/agentforce/

#Agentforce #SalesforceAI #DigitalTransformation #AchievaAI #AIAgents

Salesforce Agentforce Consulting Services

Mark Gadala-Maria (@markgadala)

한 사용자가 AI 에이전트로 '한 회사(full company)'를 구성한 전체 시스템을 오픈소스로 공개했습니다. 공개된 레포(agency-agents)를 통해 누구나 즉시 에이전시(agency)를 AI 에이전트를 이용해 운영해볼 수 있게 되었음을 알립니다. 에이전트 기반 비즈니스 운영의 실용적 데모/참고 프로젝트로 의미가 큽니다.

https://x.com/markgadala/status/2032120373114814685

#opensource #aiagents #agencyagents #github

Solo changed how I work with terminals.

https://eshlox.net/solo-changed-how-i-work-with-terminals

#Solo #DevTools #Terminal #AIAgents #DeveloperWorkflow

If your AI agent fetches web pages directly, every piece of that content flows into the model's context window unsanitized, and attackers can embed instructions your agent can't distinguish from yours. I built MCP-Airlock, a three-layer defense that runs in a separate container and catches structural, adversarial, and semantic injection attacks. Open source, AGPL, security researchers welcome. https://crunchtools.com/mcp-airlock-open-source-defense-prompt-injection-ai-agents/ #AIAgents #Security #OpenSource #PromptInjection #MCP