Denoise · Twitter

A consensus is forming between major labs like @GoogleDeepMind, @AnthropicAI, and practitioners like @MalwareTechBlog to proactively map and mitigate agent attack surfaces as they emerge.

Discussion centers on red-teaming frameworks and responsible disclosure for the new class of autonomous agents.

• 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 release of Claude Code 1.5 by @AnthropicAI, validated by @karpathy's analysis and adopted by developers like @levelsio, signals a convergence on the terminal as the new AI-native developer environment.

Major releases focus on terminal-native coding agents, with benchmarks and developer adoption stories validating the shift away from traditional 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.

A new agent stack is rapidly forming, with @OpenAI providing SDKs, platforms like @vercel and @replit offering deployment targets, and tools like @LangChainAI bridging the protocols.

Infrastructure providers and framework maintainers are shipping primitives for deploying, orchestrating, and integrating 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.

@MistralAI continues its strategy of fueling the open-source ecosystem by releasing foundational data artifacts, this time targeting vision-language capabilities.

A large-scale open dataset for Optical Character Recognition (OCR) has been released to facilitate model training.

• Mistral AI@MistralAIrising

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

With massive context windows now a reality, voices like @GregKamradt and tools like @mem0ai are moving past vector search to solve harder problems of memory management and cache invalidation.

The conversation is evolving from simple RAG to 'context engineering,' focusing on advanced memory architectures for agents in large context windows.

• 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.

Agent-like automation is becoming a standard feature in SaaS, with @NotionHQ and @linear launching capabilities for auto-filling and auto-triaging, parallel to agent orchestration from @temporalio.

Workspace productivity tools are shipping significant automation features, mirroring the agent-based workflows seen in developer tools.

• 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.

Organizations like @weights_biases are industrializing prompt optimization, running tens of thousands of variants to find efficient frontiers, a more rigorous approach than anecdotal 'tricks' from individuals.

The focus is on moving prompt engineering from an art to a science through large-scale, systematic 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.

@jerryjliu0's point on filtering synthetic data reveals a key challenge in scaling agent training: ensuring data quality to prevent models from learning spurious correlations.

A niche but critical aspect of agent infrastructure—curating high-quality training data—was highlighted.

• Jerry Liu@jerryjliu0repeated

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