Denoise · Twitter

This reveals a consensus that agent security, particularly at the orchestration layer discussed by @AlexAlbert__, is the new frontier of vulnerability research.

Major labs are publishing red-teaming frameworks and disclosures for autonomous agents, while independent researchers are testing these agents in real-world pentesting scenarios.

• Anthropic@AnthropicAIrising

  Responsible disclosure on a Claude jailbreak chain we patched last week. Full write-up including our red team timeline.

• Google DeepMind@GoogleDeepMindrising

  New red team framework for prompt injection in autonomous agents. Covers cross-tool leakage, scanner evasion, and sandbox escape patterns.

• MalwareTech@MalwareTechBlogrepeated

  Autonomous agent running pentest flows against a real SaaS. First real-world run: fewer false positives than I expected on the vulnerability surface.

The releases from @AnthropicAI and analysis from @karpathy signal a divergence from IDE-centric assistants toward stateful, terminal-based agents as the primary coding interface.

Anthropic released a terminal-native coding agent, prompting discussion and benchmarks comparing it to existing tools and highlighting a potential shift in developer workflows.

• Anthropic@AnthropicAIrising

  Claude Code 1.5 is live. Terminal-native coding agent with full Claude Opus reasoning, file-ops sandbox, and session replay.

• Andrej Karpathy@karpathyrising

  The developer-experience shift from IDE to terminal agent is underrated. Coding workflows are about to look nothing like 2024.

• swyx@swyxrising

  Codex vs Claude Code terminal agent benchmarks. Pass@1 diverges more than I expected on the long-context editor tasks.

• DSPy@dspy_airising

  DSPy 3.0: prompt optimization via compile-time search over system prompt variations. Benchmarks inside.

• @levelsio@levelsiorising

  Switched my whole editor setup to Claude Code this week. Shipping faster than when I used Cursor + Copilot.

A clear convergence is happening as major players from labs (@OpenAI) to infra (@vercel) race to provide the essential primitives for orchestrating and deploying agents.

OpenAI released an agent SDK, infrastructure providers like Vercel and Replit launched agent deployment solutions, and frameworks like LangChain added support for agent protocols.

• OpenAI@OpenAIrising

  New agent SDK: protocol-level tool calling, deployment harness, and multi-worker orchestration primitives. Docs live.

• LangChain@LangChainAIrising

  MCP protocol integration thread. How to wire existing LangGraph agents into the Anthropic Model Context Protocol server spec.

• Vercel@vercelrising

  Edge runtime for agent workers is live. Spawn durable background agents from any serverless deployment.

• Alex Albert@AlexAlbert__rising

  When your security scanner finds nothing scary on an agent deploy, check the orchestration layer again. That's usually where the jailbreak sneaks through.

• Replit@replitrising

  New agent deployment harness. One command to go from local orchestration to hosted agent worker.

While the agent conversation dominates, @MistralAI's contribution shows that foundational dataset creation remains a key differentiator for open-source model development.

Mistral AI released a large-scale open dataset for web OCR, providing a new resource for training multimodal models.

• Mistral AI@MistralAIrising

  Open dataset release: 100M-row web OCR dataset. Cleaned, licensed, ready to train.

The limitations of simple vector retrieval are becoming apparent; actors like @GregKamradt and @mem0ai are pushing for stateful, multi-layered memory systems.

The discussion moves beyond basic RAG, with experts proposing new 'context engineering' frameworks and more sophisticated memory architectures for agents.

• Vaibhav Srivastav@reach_vbrising

  Tested the new 10M context memory window end to end. Surprising failure modes around rag retrieval cache invalidation, thread below.

• Greg Kamradt@GregKamradtrising

  RAG is dead, long live context engineering. My framework for when to cache, when to retrieve, and when to just dump memory into the prompt.

• mem0@mem0airising

  Memory layer for agents: differentiating working memory from the subconscious store. Vector index isn't enough anymore.

• LlamaIndex@llamaindexrepeated

  Knowledge graph retrieval walkthrough: when semantic vector search misses, graph hop beats it every time.

The pattern of 'orchestration' from SaaS tools like @NotionHQ and workflow engines like @temporalio mirrors the agent-centric trends in AI, suggesting a broader convergence.

Productivity tools like Notion and Linear are shipping workspace automation features, while Temporal highlights its orchestration engine's applicability to agents.

• Notion@NotionHQrising

  Notion workspace automation is out of beta. Auto-fill tables, chained updates across databases, and a new audit log surface.

• Linear@linearrising

  Linear now auto-triages incoming issues. Quiet launch, but already our favorite workspace feature of the year.

• Temporal@temporaliorepeated

  Orchestrating agents with durable workflows: replayable, resumable, and multi-worker by default. Walkthrough from our infra team.

• James Clear@jamesclearrepeated

  The best habit tracker is the one you actually open. Three open-source alternatives worth trying.

The focus is shifting from prompt 'hacks' to systematic, data-driven optimization, with orgs like @weights_biases providing frameworks for scalable benchmarking.

Practitioners shared insights from large-scale system prompt benchmarking and reviews of production prompts, focusing on practical optimization techniques.

• dotey@doteyrising

  Five prompt tricks learned this week from reviewing 200 production prompts. Short thread.

• Weights & Biases@weights_biasesrising

  System prompt benchmarking at scale: we ran 40k variants across 6 frontier models. The efficient frontier is not where you think.

As agent training becomes more common, @jerryjliu0's post highlights that the core ML challenge is now data quality and curation, not just scaling compute.

A researcher shared techniques for curating synthetic datasets for agent training, focusing on filtering data that can harm model generalization.

• Jerry Liu@jerryjliu0repeated

  Dataset curation for agent training: how we filter synthetic data that looks good but poisons generalization.