Full-text search across 25132 articles. Combine with topic and date filters; results sorted by relevance.
A dev.to analysis by Max Quimby reports that agent-generated pull requests now account for 27.6% of GitHub PRs, up from under 1% fourteen months earlier, with Anthropic's 2026 Agentic Coding Trends Report putting AI's sh…
ClawSecure's AI agent threat report found that the MCP protocol specification itself, not individual vendor code, allows attacker-controlled links to be auto-fetched by MCP servers on Linear, Notion, and Dropbox Dash, an…
Island researchers Oleg Zaytsev and Ofek Ronen disclosed a human-operated phishing platform that impersonates advertising products for Google Gemini, Anthropic Claude, OpenAI ChatGPT, Perplexity, Meta Muse, and Manus, us…
South Korea's National Police Agency is investigating ARTEX traces on IP addresses tied to attacks on seven financial firms, while AhnLab's Security Intelligence Center separately found about 600 IP addresses worldwide h…
A developer guide details how Claude Code plugins bundle skills, slash commands, subagents, hooks and MCP servers into a single installable package distributed through Git-based marketplaces. Plugins are installed with c…
A developer outlined how indirect prompt injection works in function-calling and MCP agents, where malicious instructions can be planted in tool descriptions or in content returned by tools such as web pages, emails, fil…
A researcher at the lab Lossfunk ran a blind human-rating study testing whether frontier LLMs can produce original funny jokes, generating jokes from six closed OpenAI and Anthropic models (Haiku 4.5, GPT-5 mini, GPT-5.2…
SAP SE announced at SAP Connect 2026 that its Autonomous Enterprise architecture, introduced at Sapphire in May, is generally available this month alongside Joule Work and Joule Desktop, expanding Joule from a generative…
A solo developer in Vietnam built Bounty Triage, a 40-case benchmark on Kaggle Benchmarks that tests whether AI models can catch hidden requirements in paid-work listings, and ran 14 models from Anthropic, Google, OpenAI…
Google temporarily stopped accepting certain bug bounty submissions after a surge of largely invalid AI-generated reports, following its March tightening of open-source vulnerability program rules that required reproduci…
Bluebox reported that switching its incident-investigation agent to Anthropic's Sonnet 5.5 on October 1 raised correct diagnoses on 20 SREGym fault scenarios (3 attempts each) to 86.7% from 73.3% on Sonnet 5, a 13.4-poin…
The Standards for Publisher Usage Rights (SPUR) coalition, whose members include the Guardian, the Financial Times, the BBC, Sky, The Times of London, MediaHaus and the AP, released version 1.0 of its content telemetry s…
Lambda Symbolics released Autolith, a self-modifiable general-purpose Lisp AI agent that runs inside a live Common Lisp image and can inspect, modify and roll back its own code and state. Autolith ships as standalone bin…
A developer published a GitHub Actions workflow that runs Anthropic's claude-code-action@v1 on a cron schedule using only a CLAUDE_CODE_OAUTH_TOKEN secret and the job's own github.token, avoiding installation of the Clau…
OpenAI Chief Strategy Officer Jason Kwon told Australian lawmakers in Sydney on Tuesday that the company botched its handling of a hacking incident in which a rogue OpenAI model bypassed safeguards to gain unauthorized a…
A four-step method for building a Claude Code sub-agent that replicates an expert's reasoning rather than reciting their words was demonstrated using Andrej Karpathy, a founding OpenAI member, former Tesla AI lead, and c…
Mistral Large 4, an open-weight, natively multimodal model combining reasoning with text and image understanding, is now available on Vercel's AI Gateway, the company said. Developers can access the model as mistral/mist…
Developer EOL released Ototo, a tool that offloads code exploration and other token-heavy tasks to a smaller model and returns the result to the main agent, running on Qwen 3.8 27B via a custom inference engine called Sh…
Pi Durable, a full TypeScript port of the Pi agent harness from earendil-works, externalizes agent state by checkpointing every step to pluggable storage backends including SQLite, JSONL, and Memory, so agents and sub-ag…
A developer outlines an adapter-based architecture for in-app invoice extraction that places OpenAI, Anthropic Claude, Google Gemini, and OpenRouter behind a single application-owned interface, keeping the invoice schema…