Denoise · Twitter

The focus of agent security is shifting from prompt injection in single models to vulnerabilities in the orchestration layer connecting tools.

Major labs are now publishing formal red-teaming frameworks and disclosures for autonomous agents, moving security from theory to practice.

• 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 space is polarizing around two paradigms: IDE-centric copilots versus terminal-native agents, a trend highlighted by @karpathy.

Anthropic's Claude Code 1.5 launch marks a major entry into terminal-native coding agents, with immediate benchmarks and adoption stories emerging.

• 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 de facto agent stack is converging: OpenAI/Anthropic for protocols, LangChain for abstraction, and Vercel/Replit/Temporal for durable execution.

OpenAI and Anthropic are pushing for standardized agent protocols, while infra providers like Vercel and Replit ship corresponding deployment solutions.

• 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 agents dominate today's conversation, players like Mistral AI continue to invest in foundational datasets, a parallel track for improving core capabilities.

Mistral AI released a large-scale web OCR dataset, providing a foundational resource for training multimodal models.

• Mistral AI@MistralAIrising

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

A fault-line is emerging between 'stuffing the context window' and sophisticated retrieval, with @GregKamradt and @mem0ai advocating for the latter.

The discourse on agent memory is maturing beyond simple RAG, with practitioners exploring complex caching, context engineering, and knowledge graphs.

• 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 agentic workflow pattern (trigger -> multi-step action) is consolidating not just in dev tools but also in general SaaS like Notion and Linear.

Major productivity tools like Notion and Linear are releasing workspace automation features, mirroring the agent-driven trend of autonomous task completion.

• 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 engineering is becoming more systematic, with platforms like Weights & Biases enabling large-scale experiments that challenge common assumptions.

Practitioners are sharing empirical results from large-scale system prompt benchmarking, moving from anecdotal tricks to data-driven optimization.

• 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 challenge: the quality of synthetic data, not just quantity, is the bottleneck for agent generalization.

The focus in agent training data is shifting to the curation and filtering of synthetic data to avoid performance degradation.

• Jerry Liu@jerryjliu0repeated

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