Denoise · Twitter

Major labs like @AnthropicAI and @GoogleDeepMind are moving beyond prompt injection to address system-level vulnerabilities in agent orchestration and tool use.

The focus is on establishing formal security frameworks and red-teaming practices specifically for autonomous agent systems.

• 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 tool landscape is consolidating around the terminal, with @AnthropicAI's launch and @karpathy's commentary framing it as a structural shift away from IDEs.

The release of Anthropic's Claude Code 1.5 spurred discussion on a major developer experience shift towards terminal-native agents.

• 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 convergence is visible among @OpenAI, @Vercel, and @Replit to provide the "last mile" infrastructure for production agents, shifting the bottleneck from model capability to deployment reliability.

A wave of new SDKs and infrastructure for deploying and orchestrating agents was released by major platforms.

• 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 continues its pattern of contributing foundational open-source assets, this time providing a massive, cleaned dataset for training OCR models.

The sole activity was a large-scale open dataset for web OCR released by MistralAI.

• Mistral AI@MistralAIrising

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

Practitioners like @GregKamradt and startups like @mem0ai are signaling that simple vector retrieval is insufficient for stateful agents, pushing towards more sophisticated memory patterns.

The discourse is evolving from simple RAG implementations to more complex "context engineering" and layered memory architectures 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.

A convergence pattern is emerging where SaaS tools like @NotionHQ and @linear are adopting agentic primitives for workflow automation, blurring the lines with dedicated AI agent platforms.

Workspace automation tools like Notion and Linear are releasing agent-like features for autonomous task management and triage.

• 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.

Platforms like @weights_biases are industrializing prompt engineering, enabling quantitative analysis over thousands of variants to find what actually works at scale.

The emphasis is on moving from anecdotal prompt tricks to systematic, large-scale benchmarking to discover optimal system prompts.

• 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 highlights a key bottleneck in scaling agent capabilities: the difficulty of generating high-quality synthetic data that promotes, rather than harms, generalization.

The conversation focused on the critical need for careful curation of synthetic data to avoid poisoning agent training.

• Jerry Liu@jerryjliu0repeated

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