Denoise · Twitter

A consensus is forming that agent security, especially at the orchestration layer, is distinct from model safety, as shown by @AnthropicAI and @GoogleDeepMind.

The focus is on the emerging field of AI agent security, with major labs releasing red-teaming frameworks and disclosing vulnerabilities.

• 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 developer experience is fragmenting from the traditional IDE, with @karpathy and @swyx validating the terminal agent as a primary workflow.

The release of Anthropic's Claude Code 1.5 marks a significant push towards terminal-native coding agents as a primary developer interface.

• 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 race to define the standard agent stack is underway, with @OpenAI, @vercel, and @replit all shipping orchestration and deployment primitives.

Major platforms are releasing foundational infrastructure for agent development and deployment, including SDKs and managed runtimes.

• 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.

Despite the agent focus, foundational dataset releases from players like @MistralAI continue to be a key driver for open model development.

MistralAI released a large, open web OCR dataset, providing a key resource for training multimodal models.

• Mistral AI@MistralAIrising

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

The community, led by voices like @GregKamradt and projects like @mem0ai, is converging on the idea that vector search alone is insufficient for agent memory.

The conversation on memory is evolving from simple RAG to more complex "context engineering" strategies for large context models.

• 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 orchestration patterns seen in agent development (@temporalio) are now appearing in mainstream productivity tools like @NotionHQ and @linear.

Workspace automation tools are shipping significant updates, focusing on chained actions, auto-triaging, and audit logs.

• 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 convergence between practitioners like @dotey and platforms like @weights_biases indicates that prompt optimization is becoming a formal engineering discipline.

The focus is on scaling and systematizing prompt engineering, moving from artisanal tricks to large-scale, data-driven benchmarking.

• 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.

The single tweet from @jerryjliu0 highlights a critical bottleneck: the quality of synthetic data is key for agent generalization.

Discussion centers on the nuanced challenges of data curation for agent training, particularly filtering synthetic data.

• Jerry Liu@jerryjliu0repeated

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