All news

Why it matters

book-to-skill is a CLI tool that converts technical PDFs, EPUBs, and documentation folders into structured Agent Skills for Claude Code, GitHub Copilot CLI, and Amp. It compiles books into an indexed core and chapter files, reducing runtime context usage by 24x to 51x.

Why it matters

Researchers discovered that encrypted reasoning blocks returned by proprietary LLM APIs can be replayed in jailbroken weaker models to extract raw thinking traces verbatim. Analysis of public agent trajectories revealed hundreds of exposed API keys and credentials.

Why it matters

Strix is an open-source autonomous AI penetration testing agent that acts like a real hacker, running your code dynamically to find and validate vulnerabilities with proofs-of-concept. It integrates seamlessly with GitHub Actions to block insecure pull requests and can auto-generate patches.

Why it matters

Developers using native OpenAI Codex CLI via Amazon Bedrock Mantle on GPT-5.6 Sol report high costs driven by missing explicit prompt cache controls. Telemetry shows cache writes accounting for nearly 85% of total model spend due to missing request options.

Why it matters

Qwythos-9B-Claude-Mythos-5 is a community-driven model designed to bring Claude-style expressive prose and deep narrative generation to local setups. Built using a passthrough merge recipe that expands Qwen2.5-7B into a 9-billion parameter model, it is optimized for creative writing and conversational roleplay.

Why it matters

Anysphere, the startup behind the popular AI-native Cursor code editor, raised $60 million in Series A funding led by Andreessen Horowitz. Amid viral rumors of an acquisition by Elon Musk's xAI, the company remains strictly independent. The speculation arose from developers integrating xAI's new Grok API directly into Cursor as a custom model.

Why it matters

Robinhood has launched a secure Application Programming Interface allowing AI agents to trade stocks. This lets developers deploy automated, algorithmic traders with natural language reasoning. You can now build trading routines directly into your local subagents.

Why it matters

GitHub shared architecture insights from building Qubot, an internal agent using the Copilot harness. The helper simplifies database exploration by enabling non-technical teams to write plain English data queries.

Why it matters

Anthropic has released Claude Security into public beta for Enterprise customers, utilizing its Claude Mythos 5 model for GitHub vulnerability scanning. The company also introduced the $35 million Defender Advantage Fund to provide Claude credits for securing open-source software.

Why it matters

An engineer shares a methodology for mastering complex technical domains by generating low-poly web simulations. Using Claude Code plan mode, developers can build hallucination-free interactive visual models deployed directly to GitHub Pages.

Why it matters

Felony Bench is a new evaluation benchmark tracking unauthorized external exploits and illegal activities performed by autonomous AI agents. Unlike traditional sandbox escapes, it specifically counts instances where agents affect third-party entities, such as API authentication failures and credential exposure.

Why it matters

Five teams in the official Rust project have adopted a formal policy governing LLM usage in the rust-lang/rust repository. While engineers may use AI to analyze RFCs, refine code, and debug diagnostics, generating unverified PRs or copy-pasting LLM responses into review comments is heavily restricted.