Denoise · Twitter

The focus is shifting from prompt injection against a model to system-level attacks targeting the agent's orchestration and tool-use layers, as demonstrated by @GoogleDeepMind and @AnthropicAI.

Major labs are applying formal red-teaming methodologies, previously used for models, to the entire autonomous agent stack.

• 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 consensus is forming around the terminal as the next major interface for AI developers, with @AnthropicAI, @karpathy, and @swyx all pointing to a post-IDE workflow.

The launch of Anthropic's Claude Code 1.5 spurred a broad discussion on the viability of terminal-native agents replacing 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.

OpenAI, Vercel, Replit, and LangChain are all converging on providing the infrastructure for hosting and running stateful, multi-worker agents.

A wave of new tooling for agent deployment and orchestration signals a race to build the platform layer for this new computing paradigm.

• 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.

This move by @MistralAI is a foundational data play to enable community model training, rather than an application-level product release.

Mistral AI released a large-scale, cleaned web OCR dataset for public use.

• Mistral AI@MistralAIrising

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

The conversation led by @GregKamradt, @mem0ai, and @reach_vb suggests that simple vector search is now seen as insufficient for providing agents with true memory.

Discussion shifts from the mechanics of RAG to more abstract concepts like 'context engineering' and dedicated 'memory layers' for 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.

While @NotionHQ and @linear build user-facing automation, @temporalio's post on durable workflows points to the underlying orchestration engine required to make them reliable.

Workspace automation features are quietly becoming standard in SaaS tools like Notion and Linear, leveraging agent-like capabilities.

• 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.

Frameworks like @dspy_ai and platforms like @weights_biases are converging on treating prompts as a searchable, optimizable configuration space.

Efforts are underway to move prompt engineering from a craft of individual 'tricks' to a systematic, benchmark-driven science.

• 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 contribution from @jerryjliu0 serves as a reminder that sophisticated agent architectures still depend on foundational data quality practices.

A lone tweet highlights the critical, often-overlooked challenge of dataset curation and filtering to avoid poisoning agent generalization.

• Jerry Liu@jerryjliu0repeated

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