Home · 6

Move beyond basic prompting. Learn how to structure prompts for systematic testing and build an automated evaluation harness for Claude models.

ProvenanceGuard is a new verification layer for MCP agents that prevents cross-source conflation by mapping LLM claims directly to specific tool sources.

Claude Code v2.1.290 introduces managed agent onboarding, enhanced plugin permission controls, and critical fixes for long sessions and WebFetch.

Claude Code v2.1.288 introduces UI selection APIs, cleared prompt recovery, code review controls, and key fixes for gateway structured outputs.

Ai2 has released AstaBrief 8B, an open-weights model designed to generate cited scientific reports 3.5x faster than proprietary pipelines.

Ollama v0.35.1 introduces Cloudflare Clef decision models via the /v1/systemone endpoint, capability flags, and expanded web search limits.

NVIDIA Kumo Tabular delivers single-pass tabular prediction using in-context learning, topping four major benchmarks with commercial-friendly licensing.

Olmo-core 3 shifts from FSDP to DDP, delivering a 2.7x throughput boost and scaling open MoE model training to over one trillion parameters.

Anthropic updates Claude Code with extensible Mods, dangerous command safeguards, Sonnet 5.5 advisor pairing, and MCP protocol updates.

Ollama v0.35.0 introduces the systemone endpoint, bringing structured decision models like Nimble and Tev1 to local environments.