0x705 - Actu - 08 février 2026

Value of the Information & Strength of the Argument

The podcast provides valuable insights into the current state of AI agent security, particularly the risks associated with OpenClaw. The host argues convincingly that the technology is not ready for mainstream use, citing examples of malicious plugins and the lack of proper safeguards. The argumentation is solid, drawing on recent news and expert opinions, though it is somewhat opinionated. The host also makes a good point about the need for regulation and the parallels with industrial robot safety.

Scientific Rigor, Source Quality, Title Accuracy

The host references multiple sources, including news articles and reports, but does not provide direct citations within the podcast. The description includes a list of relevant links, which adds credibility. The title accurately reflects the content, which is a news update. The host’s analysis is generally rigorous, but some claims are presented without specific evidence, and the tone is sometimes editorial.

The title accurately reflects the content, which is a news update for February 8, 2026.

7/10

The podcast provides a critical analysis of recent cybersecurity and AI news, citing multiple sources and offering expert commentary. However, the lack of formal citations and the subjective nature of some opinions slightly reduce the overall reliability.

Markers derived by PSI from the transcript: the creator did not define chapters.

• 00:00Introduction and overview of the episode
• 02:00Discussion on OpenClaw and its security issues
• 10:00Legal issues with Grok and deepfakes
• 15:00Social media bans for teens in Spain and Greece
• 20:00Microsoft's rollback of AI features in Windows 11
• 25:00Vibe coding and its impact on open source
• 30:00AWS breach and AI-assisted attacks
• 35:00n8n vulnerabilities and Microsoft Copilot in OneDrive
• 40:00Industrial robot safety and parallels to AI agents
• 45:00Claude Opus 4.6 finding zero-days and trust in AI agents

• OpenClaw (a.k.a. Moltbot) is everywhere all at once, and a disaster waiting to happen — Discussed in the context of OpenClaw's security issues.
• Moltbook social media site for AI agents had big security hole, cyber firm Wiz says — Referenced when discussing the Moltbook platform's vulnerabilities.
• MoltBot Skills exploited to distribute 400+ malware packages in days — Mentioned in the context of malicious plugins for OpenClaw.
• DIY AI bot farm OpenClaw is a security 'dumpster fire' — Referenced when discussing the overall security posture of OpenClaw.
• Detecting and Monitoring OpenClaw (clawdbot, moltbot) — Mentioned as a resource for monitoring OpenClaw activity.
• Clouds rush to deliver OpenClaw-as-a-service offerings — Discussed in the context of cloud providers offering OpenClaw services.
• A sane but extremely bull case on Clawdbot / OpenClaw — Referenced when discussing the potential of OpenClaw.
• OpenClaw: When AI Agents Get Full System Access – Revolution or Security Nightmare? — Mentioned in the context of the risks of full system access.
• It’s easy to backdoor OpenClaw, and its skills leak API keys — Discussed when highlighting the ease of backdooring OpenClaw.
• Using microvm.nix to sandbox Openclaw — Mentioned as a way to sandbox OpenClaw.
• OpenClaw Partners with VirusTotal to Secure AI Agent Skill Marketplace — Referenced when discussing the partnership with VirusTotal.
• 17% of 3rd-Party Add-Ons for OpenClaw Used in Crypto Theft and macOS Malware — Mentioned in the context of malicious plugins.
• French prosecutors raid X offices, summon Musk over Grok deepfakes — Discussed in the context of legal actions against X.
• Kevin Beaumont: “The UK’s Information Commissio…” - Cyberplace — Referenced when discussing UK investigations into Grok.
• Kevin Beaumont: “Reuters reports Grok is still …” - Cyberplace — Mentioned in the context of Grok's ongoing issues.
• Kevin Beaumont: “Elon Musk Under Investigation …” - Cyberplace — Referenced when discussing investigations into Musk.
• Spain, Greece weigh teen social media bans, drawing fury from Elon Musk — Discussed in the context of social media bans for teens.
• Vos agents IA sécurisés en -10 sec. sur Mac — Mentioned as a resource for securing AI agents on Mac.
• You won: Microsoft is walking back Windows 11’s AI overload — Discussed in the context of Microsoft rolling back AI features.
• C’est prouvé : Le vibe coding va tuer l’open source — Referenced when discussing the impact of vibe coding on open source.
• Anthropic keeps Claude ad-free — Mentioned in the context of Anthropic's decision to keep Claude ad-free.
• AWS intruder pulled off AI-assisted cloud break-in in 8 mins — Discussed in the context of AI-assisted attacks.
• n8n’s latest critical flaws bypass December fix — Mentioned in the context of n8n vulnerabilities.
• Microsoft sets Copilot agents loose on your OneDrive files — Discussed in the context of Microsoft's Copilot integration with OneDrive.
• How Industrial Robot Safety Was Written In Blood — Referenced when discussing industrial robot safety.
• Anthropic’s Claude Opus 4.6 uncovers 500 zero-day flaws in open-source code — Mentioned in the context of Claude Opus 4.6's capabilities.
• GitHub - Deso-PK/make-trust-irrelevant: Make trust irrelevant for agentic AI using kernel-enforced authority boundaries. — Discussed in the context of trust in AI agents.
• Malicious VS Code AI Extensions Harvesting Code from 1.5M Devs — Mentioned in the context of malicious AI extensions.
• Notepad++ Hack Detailed Along With the IoCs and Custom Malware Used — Discussed in the context of the Notepad++ hack.
• Notepad++ Users, You May Have Been Hacked by China — Mentioned in the context of the Notepad++ hack.
• Energy infrastructure cyberattacks are suddenly in fashion — Discussed in the context of energy infrastructure attacks.
• EDR killer tool uses signed kernel driver from forensic software — Mentioned in the context of EDR bypasses.
• Microsoft releases urgent Office patch. Russian-state hackers pounce. — Discussed in the context of Microsoft Office patch.
• Attackers Using DNS TXT Records in ClickFix Script to Execute Powershell Commands — Mentioned in the context of ClickFix attacks.
• nmapUnleashed Makes Nmap Scanning More Comfortable and Effective — Mentioned as a tool for Nmap scanning.
• Google Looker Bugs Allow Cross-Tenant RCE, Data Exfil — Discussed in the context of Google Looker vulnerabilities.
• When Cloud Outages Ripple Across the Internet — Mentioned in the context of cloud outages.
• Microsoft rolls out native Sysmon monitoring in Windows 11 — Discussed in the context of Sysmon in Windows 11.
• Ukraine tightens controls on Starlink terminals to counter Russian drones — Mentioned in the context of Starlink controls.
• Satya Nadella decides Microsoft needs a quality czar — Discussed in the context of Microsoft's quality issues.
• EDR, Email, and SASE Miss This Entire Class of Browser Attacks — Mentioned in the context of browser attacks.
• GDPR is a failure — Discussed in the context of GDPR's effectiveness.
• California city turns off Flock cameras after company shared data without authorization — Mentioned in the context of Flock camera privacy issues.
• Vos données sont déjà en vente… et vous ne vous en rendez même pas compte — Discussed in the context of data selling.
• Lockdown Mode - La fonction d’Apple qui a mis le FBI en échec — Mentioned in the context of Apple's Lockdown Mode.
• We had sex in a Chinese hotel, then found we had been broadcast to thousands — Discussed in the context of privacy violations.
• Europe shrugs off tariffs, plots to end tech reliance on US — Mentioned in the context of European tech sovereignty.
• Russian spy satellites have intercepted EU communications satellites — Discussed in the context of satellite interception.
• Munich makes digital sovereignty measurable with its own score — Mentioned in the context of digital sovereignty.
• Commission trials European open source communications software — Discussed in the context of European open source software.
• Why This Computer Scientist Says All Cryptocurrency Should “Die in a Fire” — Mentioned in the context of cryptocurrency criticism.
• Bitcoin gets a zero price target in wake of Burry warning (BTC-USD:Cryptocurrency) — Discussed in the context of Bitcoin price predictions.
• Bitcoin de la “marde” ou de l’or en barre !! :) (Franck Desert) — Mentioned in the context of Bitcoin opinions.
• Flock CEO calls Deflock a “terrorist organization” — Discussed in the context of Flock's response to Deflock.
• Germany warns of Signal account hijacking targeting senior figures — Mentioned in the context of Signal hijacking.
• BrianKrebs: “Must-read: How ‘Pink Slime’ Pu…” - Infosec Exchange — Discussed in the context of pink slime journalism.
• We moved fast and broke things. It’s time for a change. — Mentioned in the context of tech industry reflection.

• OpenClaw (a.k.a. Moltbot) is everywhere all at once, and a disaster waiting to happen — Aligns with the host's critical view of OpenClaw.
• MoltBot Skills exploited to distribute 400+ malware packages in days — Supports the claim of widespread malicious plugins.
• It’s easy to backdoor OpenClaw, and its skills leak API keys — Confirms the security vulnerabilities discussed.

• A sane but extremely bull case on Clawdbot / OpenClaw — Presents a more optimistic view of OpenClaw's potential, contrasting with the host's caution.