AgentSec(A2S)

39 posts

AgentSec(A2S) banner
AgentSec(A2S)

AgentSec(A2S)

@DTINetwork_AI

Built A2S: 6-layer security framework for AI (https://t.co/Hr946LeuEL). Securing Web3 & turning threat intel into community-owned assets via AI + blockchain.

Katılım Eylül 2025
18 Takip Edilen17 Takipçiler
AgentSec(A2S)
AgentSec(A2S)@DTINetwork_AI·
🦞Shipping an open-source security audit Agent Skill for OpenClaw. 7 attack surfaces. 13 detection rules. Zero modifications. It catches what you can't see — invisible Unicode injection, memory poisoning, credential leaks, malicious skills hiding in plain text. link: github.com/AgentSec-A2S/o… Find more explanations here: #openclaw-doctor" target="_blank" rel="nofollow noopener">agentsec.vercel.app/#openclaw-doct… #AgentSec #AISecurity #A2S #OpenClaw #MoltBook #AIAgents #PromptInjection #CVE #MCP
AgentSec(A2S) tweet media
AgentSec(A2S)@DTINetwork_AI

🦞 1/ OpenClaw hit ~209K GitHub stars (as of Feb 19, 2026). Within weeks, researchers reported: • 40,214 internet‑exposed instances • ~1.5M tokens exposed via MoltBook • ~900 malicious skills flagged • 1‑click RCE: CVE‑2026‑25253 We did a full security teardown. Here’s what we found 👇 2/ We used A2S (AgentSec Stack) — 6 layers to assess agent security: L1 Identity · L2 Cognition · L3 Orchestration · L4 Action · L5 Enforcement · L6 Evidence 3/ OpenClaw shows gaps in almost every layer. But one is catastrophic: L5 — Enforcement. 4/ L5 is the non‑bypassable, fail‑closed gate between what the model decides and what the system does. Without it, “agent safety” becomes wishful thinking. 5/ The results (Wiz / SecurityScorecard / Bitdefender / NVD): • MoltBook token exposure via broken Supabase RLS • 40K+ exposed deployments in the wild • ~900 malicious skills flagged across the ecosystem • CVE‑2026‑25253: 1 click → arbitrary command execution 6/ Full deep dive (sources + every layer + mitigations): agentsec.vercel.app/blogs-insights… #AgentSec #AISecurity #A2S #OpenClaw #MoltBook #AIAgents #PromptInjection #CVE #MCP

English
0
0
4
31
AgentSec(A2S) retweetledi
Gillian Hadfield
Gillian Hadfield@ghadfield·
NIST just launched an AI Agent Standards Initiative for identity, security, and interoperability. AI agents are becoming economic actors with zero legal infrastructure in place. We require businesses to register to operate. Why expect less of AI agents? nist.gov/news-events/ne…
English
55
259
929
110.3K
LinkLayerAI
LinkLayerAI@LinkLayerAI·
Verifiable Agent vs Black Box AI 🤔 Why no one trusts trading AI? Black box model + painted curves = zero accountability. LLA is different: ✅ Trained on real positions ✅ Verifiable execution ✅ From prediction to behavior Trust = Verifiability. #VerifiableAgent #BlackBoxAI
LinkLayerAI tweet media
English
1
1
2
30K
AgentSec(A2S)
AgentSec(A2S)@DTINetwork_AI·
@orthogonal_sh Congrats on shipping this. Curation is a real step forward, but the ecosystem still needs runtime enforcement to fail‑closed when a skill slips through. In A2S we map this to L2 (scan) + L5 (enforce) + L6 (evidence). Quick 6‑layer map: agentsec.vercel.app/blogs-insights…
English
0
0
0
11
Orthogonal (YC W26)
Orthogonal (YC W26)@orthogonal_sh·
Announcing Orthogonal Skills! Some OpenClaw skills have malware. Others have 10 duplicates. You don't know which to trust. We shipped a curated set of human-reviewed skills that actually work. Built for OpenClaw first, but works with any agent supporting skills
English
8
14
68
15.3K
Lloyd 👨‍💻
Lloyd 👨‍💻@lloydtheophilus·
2026, I am open to connecting with anyone interested in: 1.Platform engineering 2. Kubernetes 3. DevOps 4. Cloud Native 5. Data on Kubernetes 6. Linux 7. Go 8. AI Ops 9. AI Agents & Agent Architectures 10. Kubernetes Security If you want to be great in tech, don’t play with community associations!!
English
158
66
846
32.7K
Amadeo Pellicce
Amadeo Pellicce@amapel·
You can now install ANY Agent skill on @Replit We scan popular community repositories and check them for security vulnerabilities. We then ensure they will work well within Replit. The result is now Agent gets tens of thousands of new capabilities 🤯 You can also bring your own skill! Try it!
Amadeo Pellicce tweet media
English
13
14
109
12K
AgentSec(A2S)
AgentSec(A2S)@DTINetwork_AI·
@t54ai Trust/Security for Agent is definitely the future main branch for agent. We put together a 6‑layer framework (A2S:Identity→Evidence) to quickly spot security gaps in agent stacks. 5–8 min overview: agentsec.vercel.app/blogs-insights…
English
0
0
0
24
t54.ai
t54.ai@t54ai·
AI agents are already moving money — unverified and unaccountable. Today, we’re announcing our $5M seed round to build the trust layer for the agentic economy. Led by Anagram, PL Capital, and Franklin Templeton, with strategic investment from Ripple. t54.ai/seed
English
240
293
1.3K
1.3M
AgentSec(A2S)
AgentSec(A2S)@DTINetwork_AI·
🦞 1/ OpenClaw hit ~209K GitHub stars (as of Feb 19, 2026). Within weeks, researchers reported: • 40,214 internet‑exposed instances • ~1.5M tokens exposed via MoltBook • ~900 malicious skills flagged • 1‑click RCE: CVE‑2026‑25253 We did a full security teardown. Here’s what we found 👇 2/ We used A2S (AgentSec Stack) — 6 layers to assess agent security: L1 Identity · L2 Cognition · L3 Orchestration · L4 Action · L5 Enforcement · L6 Evidence 3/ OpenClaw shows gaps in almost every layer. But one is catastrophic: L5 — Enforcement. 4/ L5 is the non‑bypassable, fail‑closed gate between what the model decides and what the system does. Without it, “agent safety” becomes wishful thinking. 5/ The results (Wiz / SecurityScorecard / Bitdefender / NVD): • MoltBook token exposure via broken Supabase RLS • 40K+ exposed deployments in the wild • ~900 malicious skills flagged across the ecosystem • CVE‑2026‑25253: 1 click → arbitrary command execution 6/ Full deep dive (sources + every layer + mitigations): agentsec.vercel.app/blogs-insights… #AgentSec #AISecurity #A2S #OpenClaw #MoltBook #AIAgents #PromptInjection #CVE #MCP
AgentSec(A2S) tweet media
English
1
2
3
260
AgentSec(A2S)
AgentSec(A2S)@DTINetwork_AI·
This is not just a competition; it's a glimpse into how we are all working together for the future. And we are the companions ensuring everyone moves forward on that journey! Go @DTINetwork!!
$MIA@mwa_ia

BUIDL_QUESTS 2025总决赛在新加坡完美落幕!整整一天的创新碰撞和激烈路演,为“智能体经济”注入全新活力与群体进化动力。 亮点一:权威主题演讲。Amber Group CEO @MichaelWuAmber 分享了AgentFi(智能体经济)的创新、信任和经济协作核心观点,AWS @awscloud 强调了云基础设施保障Web3项目安全可靠扩展,Maiga @Maiga_AI 聚焦人工智能驱动的区块链变革。 亮点二:10支AI创业项目路演,@AgoraSpace_ai@AiMoNetwork@AskDollyToday@BuzzingApp@DTINetwork_AI@kolect_info@LinkLayerAI@reppo@TradingFlowAI@xhunt_ai 展示了自主经济行为与多元经济激励下的创新应用。 亮点三:专业评审团阵容,涵盖Yukun(Amber.ac @amber_ac_)、Gigi(HivemindCap)、Jay Yu(PanteraCapital)、Jian(Amber Labs @ambergroup_io)、Daniel(YottaLabs)、Kenny Lai(AWS),以严谨评估和价值观对齐推动行业标准。 亮点四:冠军由AI主导的公司XHunt(@xhunt_ai)摘得,亚军为Buzzing(@BuzzingApp),季军并列AskDolly(@AskDollyToday)、TradingFlow(@TradingFlowAI)、DTINetwork(@DTINetwork_AI)。 特别感谢所有参与团队、合作伙伴与赞助商的热情贡献!群体智慧与经济协作的未来,期待与你携手共创更多高光!

English
0
0
0
105
AgentSec(A2S)
AgentSec(A2S)@DTINetwork_AI·
Together with our amazing community, we're building the next generation of security infrastructure - a collaborative defense network that adapts and grows stronger with every challenge. Thank you @mwa_ia and everyone who supported us! This is just the beginning. 🚀
English
0
0
0
41
AgentSec(A2S)
AgentSec(A2S)@DTINetwork_AI·
🎉 Exciting news! DTINetwork secured Top5 place in BUIDL_QUESTS 2025 by ai.ac @me_aiac ! This achievement reflects the growing importance of security in our rapidly evolving digital landscape. As threats become more sophisticated, community-driven protection has never been more critical.
AgentSec(A2S) tweet media
English
1
0
0
67
AgentSec(A2S)
AgentSec(A2S)@DTINetwork_AI·
Looking forward to meeting on this day!🔥 AgentFi Demo Day is a great event worth paying attention to, and we will once again share the future security defense line with everyone!!
$MIA@mwa_ia

AgentFi Demo Day by amber.ac (@amber_ac_) is approaching fast—don't blink or you'll miss the next revolution in AgentFi! Date: Sep 30, 2025 Time: 10:00–18:00 (UTC+8/GMT+8) Location: Singapore RSVP: luma.com/g6yp0sx1 Why is this event buzzing in the Agent Ecosystem? Top judges from Amber.ac, Hivemind, Pantera Capital, Amber Group, and Yotta Labs are bringing unparalleled credibility, reviewing innovations that challenge the status quo in Autonomous Financing. Powerhouse sponsors like AWS, BGA, and Rootdata highlight this isn’t your average Crypto Finance gathering—it’s a milestone for AI Startup teams ready for Population Evolution. Pitch teams to watch: Agora Space, AiMo Network, and XHunt—each pioneering new models of Economic Agent and transforming Permissionless Finance. Expect live demos, sharp panels, and a rare focus on Principle Alignment. Builders, researchers, and investors: come unlock the next wave of Economic Collaboration—this is where the future of AgentFi is born. This isn’t just an invitation—it's a challenge to every AI-Led Social Organization visionary. RSVP, drop your predictions below, and help shape the next era. See you at the core of AgentFi revolution!

English
1
0
1
108