Denoise · Twitter

The focus of agent security is shifting to the orchestration layer. Anthropic's jailbreak disclosure and DeepMind's red team framework both target multi-step, cross-tool exploits.

Major labs are publicly disclosing and building frameworks for complex, system-level agent vulnerabilities, moving beyond simple prompt injection.

• 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.

A new developer tool category is emerging, pitting Claude Code against existing setups like Copilot. Meanwhile, frameworks like DSPy are tackling agent optimization at the compiler level.

Anthropic's Claude Code 1.5 release signals a major push into terminal-native coding agents, with immediate community benchmarks and adoption.

• 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.

There is a rapid convergence on a standardized agent orchestration stack. OpenAI's SDK, LangChain's protocol integrations, and Vercel/Replit's runtimes are all building blocks for the same architectural layer.

Major infrastructure providers like OpenAI, Vercel, and Replit are releasing SDKs and runtimes for deploying and orchestrating 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.

A quiet day for this category, with the primary signal being a foundational data release from Mistral AI, which enables further model training rather than a new model itself.

Mistral AI released a large-scale, 100M-row open dataset for web optical character recognition (OCR).

• Mistral AI@MistralAIrising

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

A consensus is forming that naive RAG is insufficient. Practitioners like Greg Kamradt and companies like mem0ai are proposing more structured approaches with memory hierarchies and graph-based retrieval.

The conversation is shifting from simple RAG to 'context engineering,' exploring complex memory architectures and retrieval strategies beyond vector search.

• 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 automation features from Notion and Linear represent a domain-specific application of agentic workflows, paralleling the general-purpose agent infrastructure being built elsewhere.

Workspace automation features are becoming standard in productivity tools, with Notion and Linear launching new capabilities for auto-filling and auto-triaging.

• 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 field is bifurcating between practical heuristics from individuals like @dotey and industrial-scale analysis from platforms like @weights_biases, showing a maturation of the discipline.

Prompt engineering is evolving from sharing anecdotal tricks to systematic, large-scale benchmarking of system prompt variations.

• 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, the bottleneck is shifting from raw compute to high-quality, non-poisonous training data, a challenge highlighted by @jerryjliu0.

The focus in agent training data is on sophisticated filtering and curation to avoid synthetic data that harms model generalization.

• Jerry Liu@jerryjliu0repeated

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