DevHunt Daily
Sunday, November 1, 2026
- 1SecurityHacker News
Linux kernel CVE-2026-98432 fixed in stable releases, Ubuntu and RHEL still vulnerable
The critical use‑after‑free vulnerability in the IPv4 TCP handshake (CVE-2026-98432) was patched in several stable Linux kernel versions in October 2026. Ubuntu and Red Hat Enterprise Linux kernels remain unpatched, leaving servers exposed to remote code execution without credentials.
- 2SecurityHacker News
Molly Signal fork hasn't been updated in 3 months and remains vulnerable
The Molly Android client, a fork of Signal, last released on July 31 and still contains vulnerabilities fixed in Signal 8.27. Developers are advised to use the official Signal app instead of Molly until it receives proper updates.
- 3Coding agentsHacker News
Memory latency limits CPU utilization in database GROUP BY queries
A blog post explains that high CPU usage often masks memory latency bottlenecks, especially in GROUP BY operations that cause many cache misses. Developers should consider cache utilization to improve query performance.
- 4Dev toolsHacker News
GCC chooses jump tables or comparison trees based on switch density
The article explains how GCC decides whether to lower a switch statement into a jump table or a series of comparisons. It shows that dense case ranges produce jump tables, while sparse cases trigger comparison trees, with a threshold around eight entries per case label.
- 5SecurityHacker News
Jepsen finds ACID and reliability issues in ParticleDB 1.1.4
Jepsen testing uncovered 96 bugs in ParticleDB 1.1.4, including crashes, data loss, and transaction anomalies, showing its ACID claims were premature. The problems appear in single‑node setups and persist in later versions under default transaction settings, which matters for developers evaluating the database for AI workloads.
- 6AIHacker News
Memdoor workflow with five Qwen3.8-27B agents outperforms Fable 5 on LiveCodeBench Hard
A Memdoor example uses five Qwen3.8-27B models in distinct roles to solve programming tasks. On the first ten LiveCodeBench Hard problems it achieved an average score of 9.8/10, higher than the authors' original harness (9.6) and Fable 5 alone (8.0). Developers can examine the workflow as a reproducible multi‑agent benchmark.
- 7SecurityLobsters
Claude Code extensions for VS Code and Cursor found to contain remote access backdoor
The official Claude Code extension for VS Code and its counterpart in Cursor were discovered to include a malicious component that creates a reverse SSH backdoor on macOS, Linux and Windows. Anthropic removed the compromised version and released a fixed update, but the incident highlights a supply‑chain attack affecting many developers.
- 8Open sourceGitHub Trending
Orkas launches open-source desktop AI team with built-in specialist agents
Orkas is a free, MIT‑licensed desktop application for macOS and Windows that coordinates multiple AI agents to turn user goals into completed tasks. It includes a Commander, 16 built‑in specialist agents and supports custom keys, offering automation for research, content creation, code, and e‑commerce work.
github.com1,044 stars today - 9Open sourceGitHub Trending
DevOps-exercises repo adds over 2600 interview questions and exercises
The GitHub repository bregman-arie/devops-exercises now contains more than 2,600 questions and exercises covering topics like Linux, Jenkins, AWS, Docker, Kubernetes, Terraform and more. It is intended as a study resource for DevOps and SRE interview preparation.
github.com342 stars today - 10Open sourceGitHub Trending
Vorssaint releases free open-source macOS menu bar toolkit
Vorssaint provides a macOS menu bar app that consolidates many utilities—volume mixing, system monitoring, window management, clipboard history, and more—into a single icon. It is free, open-source, privacy‑focused, and lets users install only the features they need, which may interest developers who customize their Mac workflow.
github.com501 stars today - 11AIProduct Hunt
Agnos launches Windows app to clean documents for AI use
Agnos, a Windows‑only tool, removes repetitive headers, footers, page numbers, cookie banners and duplicates from PDFs, Word, Excel, PowerPoint and web pages. It shows each removal rule, lets users keep items, and exports cleaned text or JSONL for RAG or training data. A 14‑day free trial is followed by a one‑time fee.
producthunt.com - 12AIProduct Hunt
telemetry.dev launches LLM observability platform for AI agents
telemetry.dev provides tracing for AI agent runs, showing each model and tool call with duration, token usage, cost, and errors. It includes prompt management, evaluation against datasets, alerts, and SDKs for TypeScript and Python, helping developers debug and monitor agents.
producthunt.com - 13AIDevHunt launch
Articos MCP adds synthetic user research to test ideas within MCP agents
Articos MCP lets developers run simulated buyer tests directly in their MCP agents, supporting Claude, ChatGPT, Cursor and other clients. The tool provides feedback on headlines, landing pages, positioning, and product ideas within about 30 minutes, claiming 86% accuracy compared to human research. A free 14‑day trial with 1,000 credits is offered.
83 points - 14AIDevHunt launch
RedditAPIs launches MCP server and API alternative for AI agents
RedditAPIs provides a Reddit MCP server offering 55 tools for AI agents like Claude and Cursor, plus a separate Reddit API with 74 endpoints that bypasses Reddit's app review. Developers can access Reddit data, post, comment, and receive live updates without OAuth setup, reducing cost and latency.
73 points - 15AIDevHunt launch
OptScale Kompressor reduces coding agent output by 63% and cuts costs
OptScale Kompressor is an open‑source, deterministic tool that compresses machine‑generated output from coding agents, keeping the same bytes across turns to preserve prompt‑cache discounts. Benchmarks show a 22.7% cost reduction and 25% fewer input tokens without affecting task success rates.
24 points
