AQSACLOUDAI News

Daily AI News · November 15, 2026

AI News Today: Cursor Key Leak, Memro, Lovable Audit (Nov 15, 2026)What it means for Pakistan

By Muhammad Asif Ali · · Aqsa Cloud

In today's brief: Supply-Chain Forensics (Cursor Confirms Fireworks Key Leak), Coding Agent Memory & Reliability (Memro, Claude Code `AGENTS.md` Study), Open-Source Agent Tooling (browser-use 1.0.5, Agweb v0.3.0, TensorFold 1.2.9, Orca v1.4.259), Chinese Model De-Censoring (Hirundo & Lazarus AI), Security Audits (Lovable App Data Exposure), and Platform Incidents (DeepSeek Outage & Meta $887M Fine).

🇵🇰 Jump to: What this means for Pakistan

🔥 Top Developments

🇵🇰 What This Means for Pakistan

🧠 New AI Models

🚀 Model Updates

🔬 AI Research

💻 Coding AI

🤖 AI Agents

🎨 Image, Video & Audio AI

🌏 China & Open-Weight AI

💰 AI Business

⚠️ Leaks & Rumors

📊 Model Race

Category / Focus Current Signal Key Challenger / Tool Open-Weight / Local Note
Coding Agent Memory Memro (MCP + PostgreSQL) MemGit / Hippocampus Open-source Go service (Nov 14)
Browser Agent Safety browser-use 1.0.5 (Self-Judge) Agweb v0.3.0 (--secret-cmd) Both open-source (Nov 14)
Local Inference Serving llama.cpp v1.0.0 TensorFold 1.2.9 (Ollama API) Metal, CUDA, Vulkan, HIP support
Multi-Agent Orchestration Google Antigravity /teamwork Orca v1.4.259 (Kimi Code) Scoped file/tool access per step

⭐ Aqsa Cloud Takeaway

  1. Rotate Keys and Scrub Traces Sent to Third-Party Inference Vendors: Following Cursor's confirmation of the Fireworks AI key leak, audit what coding-agent traces or .env variables your IDE or backend sends to external inference clouds, and disable unredacted trace logging immediately.
  2. Consolidate Project Instructions into CLAUDE.md (or Verify AGENTS.md Loading): Because Claude Code skips AGENTS.md in 16% of dual-file repos, keep critical build and security rules in CLAUDE.md—or reference AGENTS.md explicitly from CLAUDE.md—so your coding agent never misses project constraints.
  3. Self-Host Memro to Stop Re-Explaining Architecture Across Sessions: Deploy Memro with a local PostgreSQL instance to share persistent project memory across Claude Code, Cursor, and Codex over MCP.
  4. Audit Vibe-Coded Apps for Public Database & Auth Exposure: In light of the Lovable Common Crawl audit uncovering thousands of publicly readable tables, always verify Row-Level Security (RLS) and authentication gates before shipping AI-scaffolded web apps.
  5. Upgrade Browser Automation to browser-use 1.0.5 or Agweb v0.3.0: Use browser-use 1.0.5's built-in prompt-injection judge and Agweb's --secret-cmd injection pattern (Munder Difflin) so your browser agents never expose login credentials inside LLM context windows.