Denoise · Twitter

A consensus is emerging that the agent orchestration layer is the primary attack surface, a view shared by both @GoogleDeepMind's framework and @MalwareTechBlog's pentesting.

Today's discussion centers on formalizing agent security through red-teaming frameworks and public disclosure of patched jailbreaks.

• 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 release, combined with @karpathy's analysis, presents a direct challenge to the Copilot/Cursor paradigm, suggesting a fundamental workflow shift.

The release of Anthropic's Claude Code 1.5 has catalyzed a strong narrative around terminal-native agents supplanting IDE-based 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.

@OpenAI, @vercel, and @replit are converging on providing a 'PaaS for agents,' abstracting the complexities of production deployment for developers.

Major platforms are releasing SDKs and deployment harnesses focused on orchestrating and hosting durable, multi-worker AI 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 dataset release directly enables smaller actors to train competitive vision models, challenging the proprietary data moats of larger labs.

The category is quiet today, dominated by a single large-scale open dataset release for web OCR.

• Mistral AI@MistralAIrising

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

A fault-line is appearing between scaling raw context (@reach_vb) and architecting explicit memory layers (@mem0ai, @GregKamradt), suggesting RAG is no longer a monolithic solution.

The conversation is shifting from simple RAG to 'context engineering,' exploring complex memory architectures and the failure modes of massive context windows.

• 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 parallel releases from @NotionHQ and @linear show a convergence in SaaS on providing autonomous, background task automation for users.

Productivity tools are shipping significant workspace automation features, moving towards agent-like background processes.

• 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.

Work by @weights_biases refutes the 'bag of tricks' approach, consolidating a view of prompt engineering as a scalable, data-driven optimization problem.

The focus is on systematic, large-scale benchmarking of prompts to find efficient frontiers, rather than artisanal prompt crafting.

• 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.

@jerryjliu0's insight implies that the quality and curation of synthetic data, not just quantity, is becoming a key differentiator in training capable agents.

A single signal highlights the critical need for sophisticated filtering of synthetic data to avoid degrading agent performance.

• Jerry Liu@jerryjliu0repeated

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