Denoise · Twitter

Agent security is emerging as a distinct subfield, with @AnthropicAI and @GoogleDeepMind publishing formal red-teaming frameworks while practitioners like @MalwareTechBlog validate them in real-world pentests.

Major labs and independent researchers are defining the new attack surface for autonomous agents, focusing on prompt injection and orchestration 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 convergence on the terminal as the new agent-native IDE is accelerating, validated by user adoption (@levelsio) and competitive benchmarking (@swyx) following the Claude Code release.

Anthropic's release of Claude Code 1.5, a terminal-native agent, marks a significant push to move the developer's primary interface from the IDE to the command line.

• 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.

The agent stack is standardizing around orchestration (OpenAI SDK), protocol integration (LangChain), and serverless execution (Vercel), creating a de-facto architecture for production agents.

Major infrastructure providers including OpenAI, Vercel, and Replit released new primitives for orchestrating, deploying, and running autonomous 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.

While agent tooling dominates today's conversation, foundational model builders like MistralAI continue to invest in the data supply chain required for the next generation of multimodal capabilities.

MistralAI released a large-scale, cleaned web OCR dataset to the public, providing a foundational asset for training future multimodal models.

• Mistral AI@MistralAIrising

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

The discourse is shifting from 'RAG' to 'context engineering', with @GregKamradt, @mem0ai, and LlamaIndex all proposing more structured approaches to memory and retrieval.

Developers are probing the limits of large context windows and evolving beyond simple RAG, developing new patterns for memory management in 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.

A convergence is visible between no-code automation (Notion, Linear) and stateful orchestration (Temporal), with both approaches targeting the need for reliable, multi-step autonomous workflows.

Workspace automation tools like Notion and Linear are shipping agent-like features, while durable execution frameworks like Temporal are being positioned for agent orchestration.

• 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 data-driven approach to prompt optimization is solidifying, with tools from @dspy_ai and @weights_biases enabling developers to programmatically search the design space instead of relying on intuition.

Prompt engineering is maturing from sharing individual tricks to systematic, large-scale benchmarking to find optimal system prompt configurations.

• 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 accessible, the focus shifts from raw compute to subtle data quality issues, with practitioners like @jerryjliu0 highlighting failure modes specific to synthetic data.

A discussion emerged around the nuances of dataset curation for training agents, specifically on filtering synthetic data to avoid harming model generalization.

• Jerry Liu@jerryjliu0repeated

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