Denoise · Twitter

The focus of security research is shifting from chatbot prompt injection to orchestration-level exploits, with @AnthropicAI and @GoogleDeepMind leading the public discourse on agent security.

Major AI labs are now publicly disclosing agent-specific vulnerabilities and releasing red-teaming frameworks for autonomous systems.

• 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.

@AnthropicAI's Claude Code is directly competing with the Copilot/IDE paradigm, a structural shift validated by influential figures like @karpathy.

The ecosystem is rapidly adopting terminal-native coding agents, with user reports and benchmarks suggesting a workflow shift away from traditional IDEs.

• 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 de facto standard for agent deployment is emerging, with @OpenAI, @vercel, and @replit building similar primitives, while @LangChainAI provides interoperability layers.

Infrastructure providers are converging on agent orchestration, releasing dedicated SDKs, deployment harnesses, and runtimes for stateful, multi-worker agents.

• 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.

@MistralAI's open dataset release signals a move to commoditize the data layer for vision-language models, fostering more open-source competition.

A large, cleaned, and licensed web OCR dataset has been released to the public, aimed at accelerating open multimodal model training.

• Mistral AI@MistralAIrising

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

The simple RAG pattern is fragmenting, with developers like @GregKamradt, @mem0ai, and @llamaindex advocating for more structured approaches like knowledge graphs and layered memory systems.

The conversation shifts from simple RAG to 'context engineering', exploring more complex memory architectures beyond basic vector retrieval.

• 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 principles of agentic workflows, visible in tools from @temporalio, are now being embedded into mainstream productivity apps like @NotionHQ and @linear.

Workspace automation is becoming a standard feature in SaaS tools, with products like Notion and Linear adding agent-like auto-triage and chained updates.

• 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.

A split is emerging between manual prompt crafting (@dotey) and industrial-scale optimization, where platforms like @weights_biases enable data-driven prompt discovery.

The discipline of prompt engineering is maturing from sharing anecdotal tricks to performing large-scale, systematic benchmarking of system prompts.

• 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 capabilities advance, experts like @jerryjliu0 highlight that the bottleneck is no longer just compute, but sophisticated data filtering to prevent model poisoning.

The focus for training advanced agents is shifting towards the challenges of high-quality synthetic data curation and filtering.

• Jerry Liu@jerryjliu0repeated

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