Denoise · Twitter

The focus of agent security is moving beyond simple prompt injection to complex exploits in the orchestration layer and cross-tool communication.

Red teams at Anthropic and Google DeepMind are publishing disclosures and frameworks for vulnerabilities in 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 primary interface for AI coding assistance is shifting from IDE extensions like Copilot to standalone terminal agents like Claude Code, with early benchmarks showing performance divergence.

Anthropic's release of Claude Code 1.5, a terminal-native agent, has triggered discussion about a fundamental shift in developer workflows.

• 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 stack for agents, with OpenAI's SDK providing the protocol and Vercel/Replit offering the managed execution environments.

OpenAI, Vercel, and Replit have all released new infrastructure for orchestrating and deploying 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 most of the focus is on agent capabilities, Mistral's data release underscores the foundational importance of high-quality, large-scale public datasets for advancing core model modalities.

Mistral AI released a 100M-row open dataset for web OCR, providing a significant new resource for training multimodal models.

• Mistral AI@MistralAIrising

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

The conversation is maturing from a binary choice between RAG and long context to designing sophisticated memory systems, with mem0ai and Greg Kamradt proposing new frameworks.

Discussions have moved beyond simple RAG, focusing on 'context engineering' with complex caching, memory hierarchies, and graph retrieval.

• 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 AI agents are transforming coding workflows, a parallel trend sees platforms like Notion and Linear embedding 'autopilot' capabilities for administrative and project management tasks.

Workspace automation is advancing, with Notion and Linear releasing features for auto-filling data and triaging issues.

• 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.

Prompt optimization is industrializing. Weights & Biases' 40k-variant benchmark shows a move towards data-driven discovery of optimal prompts, treating it as a formal hyperparameter tuning problem.

The practice of prompt engineering is becoming more systematic, with a focus on large-scale benchmarking over anecdotal tricks.

• 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 scales, the focus shifts from raw compute to data quality. Jerry Liu's point on filtering synthetic data highlights the emerging challenge of preventing model degradation from seemingly good but flawed training examples.

The conversation centers on the critical need for high-quality data curation when training agents, particularly filtering synthetic data.

• Jerry Liu@jerryjliu0repeated

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