CP SignalTechnology worth your attention.

Wednesday, 21 October 2026

The Daily Edition: 5 stories, ranked by significance

  1. Security & privacy

    8 eMudhra certificates are compromised. Here's what you need to know

    Cloudflare disclosed that private keys for eight eMudhra intermediate certificates, widely trusted by browsers, operating systems and Cloudflare itself, were compromised on October 19, 2026. Because the keys are weak, an attacker could forge valid certificates for any domain to intercept or modify TLS traffic. Cloudflare has patched its WARP/1.1.1.1 clients and cloudflared, deployed mitigations across its CDN and Zero Trust products, and is urging customers using custom trust stores, mTLS, or strict origin TLS modes to review and update their configurations.

    Read the storyNo. 1 of 5Cloudflare BlogTue 20 Oct
  2. Security & privacy

    Fake News, Real Crashes: Information Injection Attacks on LLM Financial Trading Agents

    Researchers built a benchmark testing LLM financial trading agents against fabricated news and data rather than instruction-based prompt injection, running roughly 10,000 decisions across 15 U.S. stocks. They found fabricated news and fake indicator panels were dramatically more effective than classic prompt injection at manipulating agent trades, and that surface-level credibility cues like a named source or "BREAKING" label mattered more than the content of the fake claim itself.

    No. 2 of 5arXiv cs.SEMon 19 Oct
  3. Artificial intelligence

    Claude designed protein binders against scorpion neurotoxins

    Anthropic had an internal Claude research model design protein binders against five scorpion neurotoxins, working largely autonomously over 64 hours with four human follow-up prompts. It generated 318,484 candidate designs, narrowed these to 211 that were lab-tested by Adaptyv Bio and Twist Bioscience, and 17 met Anthropic's criteria as binders against individual toxins (none succeeded against multiple toxins). Anthropic explicitly states this is not a confirmed antivenom, only a step toward one, since neutralization in a cellular environment has not been verified.

    No. 3 of 5Anthropic NewsTue 20 Oct
  4. Software development

    SWE-Residual: Evaluating Starting-State Robustness of Coding Agents

    A new arXiv benchmark, SWE-Residual, tests coding agents not from a clean repository state but from a state where a predecessor agent already attempted the task and (wrongly) reported it complete. Across five frontier agents, success dropped 20.6-53.5 percentage points on DeepSWE and 20.6-34.9 points on Terminal-Bench versus clean starts, with agent rankings reversing on DeepSWE. The authors attribute most failures to 'acceptance bias': agents treat inherited work as the definition of the remaining task rather than re-checking it against the original request.

    No. 4 of 5arXiv cs.AIMon 19 Oct
  5. Artificial intelligence

    Towards Architectural-Level Safety Control: Relocating Hazardous Knowledge into Removable Experts in Mixture-of-Experts Models

    A new paper proposes REX, a method that relocates hazardous knowledge (e.g., bio-weapons-relevant content) into dedicated, physically removable expert modules within Mixture-of-Experts models, rather than just suppressing it through weight adjustments. Tested on three MoE models, excising these experts substantially cuts hazardous-knowledge benchmark scores while barely touching general capability, and crucially the removal resists being undone by subsequent fine-tuning, unlike existing unlearning baselines.

    No. 5 of 5arXiv cs.LGMon 19 Oct