# AI Agent

An AI agent is a goal-directed software system that autonomously plans and executes tasks by dynamically using tools and evaluating results.

This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles.

## "We are not querying like humans anymore": OpenSearch's Bianca Lewis on why AI agent mistakes look like successes

DevFeed: ["We are not querying like humans anymore": OpenSearch's Bianca Lewis on why AI agent mistakes look like successes](<https://devfeed.tech/articles/we-are-not-querying-like-humans-anymore-opensearch-s-bianca-lewis-on-why-ai-agent-mistakes-look-like-successes-146621.md>)

Original publisher: [Read original article](<https://thenewstack.io/opensearch-ai-agent-observability/>)

Author: Chris J. Preimesberger

Published: 2026-11-20T19:37:08Z

Content type: news

Language: en

Sources: [The New Stack](<https://devfeed.tech/sources/the-new-stack.md>)

Topics: [opensearch](<https://devfeed.tech/topics/opensearch.md>), [AI Agent](<https://devfeed.tech/topics/ai-agent.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agentic](<https://devfeed.tech/tags/agentic.md>), [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [ai-models](<https://devfeed.tech/tags/ai-models.md>), [ai-strategy](<https://devfeed.tech/tags/ai-strategy.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [api](<https://devfeed.tech/tags/api.md>), [apis](<https://devfeed.tech/tags/apis.md>), [applications](<https://devfeed.tech/tags/applications.md>), [cncf](<https://devfeed.tech/tags/cncf.md>), [community](<https://devfeed.tech/tags/community.md>), [context](<https://devfeed.tech/tags/context.md>), [control](<https://devfeed.tech/tags/control.md>), [cost](<https://devfeed.tech/tags/cost.md>), [dashboard](<https://devfeed.tech/tags/dashboard.md>), [data-gravity](<https://devfeed.tech/tags/data-gravity.md>), [error](<https://devfeed.tech/tags/error.md>), [hosting](<https://devfeed.tech/tags/hosting.md>), [mistakes](<https://devfeed.tech/tags/mistakes.md>), [opensearch](<https://devfeed.tech/tags/opensearch.md>), [podcast](<https://devfeed.tech/tags/podcast.md>), [sponsor-cncf](<https://devfeed.tech/tags/sponsor-cncf.md>), [sponsored](<https://devfeed.tech/tags/sponsored.md>), [sponsored-podcast](<https://devfeed.tech/tags/sponsored-podcast.md>), [video](<https://devfeed.tech/tags/video.md>)

### AI overview

AI agents can make the wrong API calls yet receive successful responses, leaving ordinary error logs and green dashboards unable to reveal that a run went off course. OpenSearch Software Foundation executive director Bianca Lewis describes using OpenSearch with OpenTelemetry to trace the services and APIs an agent called, and adding cost data to spot expensive steps. Those traces help developers investigate plausible but incorrect results and reconsider the context made available to the agent. Lewis also says teams may use OpenSearch for both observability and search, drawing on existing skills and reviewing duplicated data, while weighing platform-change costs against operating costs. Moving platforms takes time because production systems, contracts and team skills persist. The excerpt also points to community contributions through pull requests, Slack discussions and user groups.

### Source excerpt

This situation is more common than one might think: An AI agent calls the wrong API, gets a perfectly good The post "We are not querying like humans anymore": OpenSearch's Bianca Lewis on why AI agent mistakes look like successes appeared first on The New Stack.

## Best platforms using LangGraph for stateful agent workflows

DevFeed: [Best platforms using LangGraph for stateful agent workflows](<https://devfeed.tech/articles/best-platforms-using-langgraph-for-stateful-agent-workflows-146332.md>)

Original publisher: [Read original article](<https://www.lyzr.ai/blog/best-platforms-using-langgraph-for-stateful-agent-workflows/>)

Author: Sujatha R

Published: 2026-11-20T13:14:16Z

Content type: comparison

Language: en

Sources: [Lyzr.ai](<https://devfeed.tech/sources/lyzr-ai.md>)

Topics: [Langgraph](<https://devfeed.tech/topics/langgraph.md>), [Stateful](<https://devfeed.tech/topics/stateful.md>), [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [agent observability](<https://devfeed.tech/topics/agent-observability.md>), [React](<https://devfeed.tech/topics/react.md>)

Tags: [a2a](<https://devfeed.tech/tags/a2a.md>), [a2a-protocol](<https://devfeed.tech/tags/a2a-protocol.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agent-studio](<https://devfeed.tech/tags/agent-studio.md>), [agent-workflows](<https://devfeed.tech/tags/agent-workflows.md>), [agentcore](<https://devfeed.tech/tags/agentcore.md>), [agentic-workflow](<https://devfeed.tech/tags/agentic-workflow.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent-memory-systems](<https://devfeed.tech/tags/ai-agent-memory-systems.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [ai-guardrails](<https://devfeed.tech/tags/ai-guardrails.md>), [air-gapped](<https://devfeed.tech/tags/air-gapped.md>), [durable-execution](<https://devfeed.tech/tags/durable-execution.md>), [enterprise](<https://devfeed.tech/tags/enterprise.md>), [governance](<https://devfeed.tech/tags/governance.md>), [langgraph](<https://devfeed.tech/tags/langgraph.md>), [memory](<https://devfeed.tech/tags/memory.md>), [platforms](<https://devfeed.tech/tags/platforms.md>), [stateful](<https://devfeed.tech/tags/stateful.md>), [workflows](<https://devfeed.tech/tags/workflows.md>)

### AI overview

LangGraph represents agent workflows as graphs with shared state and checkpoints, allowing runs to pause, retry, or recover without restarting. The text compares five platforms: Lyzr connects existing LangGraph agents over A2A and adds SuperFlow retries, approval gates, audit trails, and private or air-gapped deployment; LangSmith Agent Server provides managed durable execution, checkpoints, queues, streaming, and cloud, hybrid, or self-hosted options, with potentially metered costs; Amazon Bedrock AgentCore offers AWS session and long-term memory, identity, Cedar policies, and telemetry, but its many separately billed components may complicate forecasting; Databricks Mosaic AI uses Lakebase for checkpoints and semantic memory under Unity Catalog governance, with setup and cost complexity for teams without Spark or MLOps experience; CopilotKit synchronizes LangGraph state with React and Next.js interfaces for live updates and approvals, while backend isolation and observability may remain the engineering team's responsibility. The Lyzr integration requires a deployed A2A endpoint, credentials when authentication is enabled, and Agent Studio access; its setup imports the agent, adds

### Source excerpt

Compare the best platforms using LangGraph for stateful agent workflows, from durable execution and memory to enterprise governance. Your support agent starts a refund request,...

## Five Guardrails That Limit the Damage an AI Agent Can Do

DevFeed: [Five Guardrails That Limit the Damage an AI Agent Can Do](<https://devfeed.tech/articles/five-guardrails-that-limit-the-damage-an-ai-agent-can-do-145977.md>)

Original publisher: [Read original article](<https://towardsdatascience.com/five-guardrails-that-limit-the-damage-an-ai-agent-can-do/>)

Author: Thomas Reid

Published: 2026-11-20T12:00:00Z

Content type: tutorial

Language: en

Sources: [Towards Data Science](<https://devfeed.tech/sources/towards-data-science.md>)

Topics: [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [Security](<https://devfeed.tech/topics/security.md>), [function calling](<https://devfeed.tech/topics/function-calling.md>), [Test-driven development](<https://devfeed.tech/topics/tdd.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [arguments](<https://devfeed.tech/tags/arguments.md>), [guardrail](<https://devfeed.tech/tags/guardrail.md>), [guardrails](<https://devfeed.tech/tags/guardrails.md>), [llmops](<https://devfeed.tech/tags/llmops.md>), [pydantic](<https://devfeed.tech/tags/pydantic.md>), [pytest](<https://devfeed.tech/tags/pytest.md>), [python](<https://devfeed.tech/tags/python.md>), [sqlite](<https://devfeed.tech/tags/sqlite.md>), [tests](<https://devfeed.tech/tags/tests.md>)

### AI overview

The article's central safeguard is to let Python enforce what an AI agent can do, since prompts alone cannot reliably prevent harmful actions. It treats user questions, model tool calls, and returned data as untrusted, including indirect prompt injection in database notes. The example uses Pydantic models to reject unknown fields and constrain arguments, an allowlisted dispatcher for read tools, read-only SQLite connections with limited query results, and a cap on the agent's turns. A return write stays outside the autonomous tool loop and is saved only after user confirmation; a UUID idempotency key and database uniqueness constraint prevent duplicate requests on retry. Pytest checks the deterministic controls. The example agent ignored Mina's malicious note, but the article cautions that this does not fully mitigate prompt injection; application-enforced limits reduce the damage a bad model decision can cause.

### Source excerpt

AI Agents for Beginners, Part three: strict Pydantic arguments, a tool allowlist, read-only SQLite, a step cap, and confirmed writes, all verified by five pytest tests in 0.68s. The post Five Guardrails That Limit the Damage an AI Agent Can Do appeared first on Towards Data Science.

## Apple doesn't need the smartest AI agent. It needs the most trusted one

DevFeed: [Apple doesn't need the smartest AI agent. It needs the most trusted one](<https://devfeed.tech/articles/apple-doesn-t-need-the-smartest-ai-agent-it-needs-the-most-trusted-one-146557.md>)

Original publisher: [Read original article](<https://www.computerworld.com/article/4251496/apple-doesnt-need-the-smartest-ai-agent-it-needs-the-most-trusted-one.html>)

Author: Jonny Evans

Published: 2026-11-20T10:30:00Z

Content type: opinion

Language: en

Sources: [Computerworld](<https://devfeed.tech/sources/computerworld.md>)

Topics: [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [agentic workflows](<https://devfeed.tech/topics/agentic-workflows.md>), [Multi-step AI processes](<https://devfeed.tech/topics/multi-step-ai-processes.md>), [Local AI](<https://devfeed.tech/topics/local-ai.md>), [least privilege](<https://devfeed.tech/topics/least-privilege.md>), [ai security](<https://devfeed.tech/topics/ai-security.md>)

Tags: [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [apple](<https://devfeed.tech/tags/apple.md>), [apple-artificial-intelligence-productivity-software-siri-vendors-and-providers-voice-assistants](<https://devfeed.tech/tags/apple-artificial-intelligence-productivity-software-siri-vendors-and-providers-voice-assistants.md>), [artificial-intelligence](<https://devfeed.tech/tags/artificial-intelligence.md>), [edition-us](<https://devfeed.tech/tags/edition-us.md>), [language-en](<https://devfeed.tech/tags/language-en.md>), [privacy](<https://devfeed.tech/tags/privacy.md>), [private-cloud-compute](<https://devfeed.tech/tags/private-cloud-compute.md>), [productivity](<https://devfeed.tech/tags/productivity.md>), [productivity-software](<https://devfeed.tech/tags/productivity-software.md>), [siri](<https://devfeed.tech/tags/siri.md>), [siri-ai](<https://devfeed.tech/tags/siri-ai.md>), [storytype-opinion](<https://devfeed.tech/tags/storytype-opinion.md>), [vendors-and-providers](<https://devfeed.tech/tags/vendors-and-providers.md>)

### AI overview

The article argues Apple's advantage in agentic AI could come from trust, integration and distribution rather than having the smartest model. It describes safeguards already in place: users are prompted to consider risks before granting agents Full Disk Access, and agents can be confined to specified folders, while Siri provides context for actions. The author expects demanding tasks to use Private Cloud Compute and speculates Apple may add Secure Exclave to minimize uploaded information; local Macs could handle some AI analysis. Whether this approach succeeds depends on reliable execution of complex, multistep actions, including how well App Intents perform. These are partly predictions, not confirmed features or outcomes.

### Source excerpt

When MacPaw, Proton, Google, OpenAI, Anthropic, and others are doing it, it's not really such a surprise that Apple is working to introduce agentic AI tools in Siri and Apple Intelligence. The wider question is how the company plans to deploy these tools -- how it will deliver the maximum convenience at the least possible cost to user privacy and data integrity. Some of those answers are already easy to speculate on. Apple has already put some of the building blocks in place, including the recent move to make consumers think more deeply when providing AI agents with Full Disk Access. Convenience versus control I know some competitors are complaining about this, but frankly I think user privacy outweighs commercial convenience -- users should have a conscious choice, which Apple delivers. "Addressing this is critical," Apple told developers. "As AI agents become increasingly capable and autonomous, the risks associated with this level of access will grow substantially. We are committed to ensuring users clearly understand these risks before granting such access, so they can make informed decisions about their own data and privacy." Apple's decision to make it possible to confine agents to folder access also speaks volumes, as it shows us how it will probably choose to manage its own agents. Siri AI may have contextual understanding of everything, which can inform Siri Agents, but the agents themselves will have but limited access to your data. They gain contextual direction from the user and from Siri AI but are limited to the proscribed folders and files when it comes to acting. Making the cloud sweat a little Some tasks are harder than others. The really hard tasks will still lean into Apple's Private Cloud Compute, and I think the company will try to introduce Secure Exclave into that relationship, so even the most complex task instructions only upload the minimum quantity of your information required to get a task done. Apple sells hardware, of course, so it makes

## Best Rate Limiting and Quota Tools for Agent Fleets

DevFeed: [Best Rate Limiting and Quota Tools for Agent Fleets](<https://devfeed.tech/articles/best-rate-limiting-and-quota-tools-for-agent-fleets-146955.md>)

Original publisher: [Read original article](<https://www.swytchcode.com/content/best-rate-limiting-quota-tools>)

Author: Swytchcode Team

Published: 2026-11-20T09:00:00Z

Content type: comparison

Language: en

Sources: [Swytchcode](<https://devfeed.tech/sources/swytchcode.md>)

Topics: [rate-limiting](<https://devfeed.tech/topics/rate-limiting.md>), [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [Tool](<https://devfeed.tech/topics/tool.md>), [Multi Agent Systems](<https://devfeed.tech/topics/multi-agent-systems.md>)

Tags: [agent-fleets](<https://devfeed.tech/tags/agent-fleets.md>), [api](<https://devfeed.tech/tags/api.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [quota](<https://devfeed.tech/tags/quota.md>), [rate-limiting](<https://devfeed.tech/tags/rate-limiting.md>), [tools](<https://devfeed.tech/tags/tools.md>)

### AI overview

Agent fleets need a shared counter to stay within a provider quota: per-process token buckets cannot see calls from other workers. The article compares shared-counter options: Kong can use Redis, while its local policy only approximates a fleet limit; Tyk's shared multi-node behavior depends on licensed components. Unkey shares customer limits across keys, with regional counts that may run high. Upstash Ratelimit uses Redis over HTTP for short-lived workers. Envoy delegates global decisions to an external rate-limit service, typically backed by Redis, which adds operational work. Treblle provides monitoring and alerts but does not throttle traffic. A Redis limiter built in-house offers control but requires atomic operations and ownership of edge cases. These tools address quota pooling, while safe retries are separate: the article says Swytchcode retries rate-limited calls with backoff and automatically retries POST or PATCH only when an idempotency key is set.

### Source excerpt

Most api rate limiting tools protect one process. Agent fleets share one provider quota, and that's a different problem.

## Best Enterprise AI Agent Platforms in 2026

DevFeed: [Best Enterprise AI Agent Platforms in 2026](<https://devfeed.tech/articles/best-enterprise-ai-agent-platforms-in-2026-146951.md>)

Original publisher: [Read original article](<https://www.swytchcode.com/content/best-enterprise-ai-agent-platforms>)

Author: Swytchcode Team

Published: 2026-11-20T09:00:00Z

Content type: comparison

Language: en

Sources: [Swytchcode](<https://devfeed.tech/sources/swytchcode.md>)

Topics: [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [agent orchestration](<https://devfeed.tech/topics/agent-orchestration.md>), [Multi Agent Systems](<https://devfeed.tech/topics/multi-agent-systems.md>), [agent observability](<https://devfeed.tech/topics/agent-observability.md>)

Tags: [2026](<https://devfeed.tech/tags/2026.md>), [agent](<https://devfeed.tech/tags/agent.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [best](<https://devfeed.tech/tags/best.md>), [enterprise](<https://devfeed.tech/tags/enterprise.md>), [enterprise-ai](<https://devfeed.tech/tags/enterprise-ai.md>), [frameworks](<https://devfeed.tech/tags/frameworks.md>), [platforms](<https://devfeed.tech/tags/platforms.md>), [runtimes](<https://devfeed.tech/tags/runtimes.md>)

### AI overview

The article argues that enterprise AI agent platforms serve different stages, so a single ranking obscures their roles. LangGraph, CrewAI and OpenAI Agents SDK handle agent reasoning; Composio, Arcade, Pipedream and Zapier MCP make tools reachable, with differences in OAuth, deployment, pricing and production access. Workato is positioned for mostly deterministic integration flows. The article describes Swytchcode as a production runtime that governs calls after an agent chooses them: it checks schemas and policies, injects credentials at execution, validates responses, audits calls, and retries POST/PATCH only when an idempotency key is set. It says catalogs' public docs do not describe write idempotency as of November 2026. Swytchcode is presented for costly or sensitive writes, but it lacks a hosted branded OAuth connection flow for large numbers of end users; the article suggests pairing it with a catalog when both needs apply. The text is a vendor-authored comparison that includes Swytchcode among the nine platforms.

### Source excerpt

We score the real enterprise AI agent platforms (frameworks, catalogs, iPaaS, runtimes) against the seven stages where agent builds actually break.

## CompAct-Bench: A Benchmark for Working-Context Compaction in Long-Horizon Agent Tasks

DevFeed: [CompAct-Bench: A Benchmark for Working-Context Compaction in Long-Horizon Agent Tasks](<https://devfeed.tech/articles/compact-bench-a-benchmark-for-working-context-compaction-in-long-horizon-agent-tasks-145569.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.25673>)

Author: Jiayi Chen, Lizhe Chen, Yujia Chen, Shihan Dou, Ming Zhang, Chendi Ge, Shiguang Ni, Tao Gui, Qi Zhang, Xuanjing Huang, Maxm Pan

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [Frontier AI](<https://devfeed.tech/topics/frontier-ai.md>), [Claude](<https://devfeed.tech/topics/claude.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agent-tasks](<https://devfeed.tech/tags/agent-tasks.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [context](<https://devfeed.tech/tags/context.md>), [context-compaction](<https://devfeed.tech/tags/context-compaction.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [software-engineering](<https://devfeed.tech/tags/software-engineering.md>)

### AI overview

CompAct-Bench measures whether an agent can continue a task after its working context is compressed. It inserts compaction points into successful trajectories, has the model under evaluation compress the preceding context, then lets the same executor resume; compactors are compared at controlled retention ratios. The benchmark spans BrowseComp, SWE-bench, and an internal CompanyBench evaluation. The abstract reports that Claude- and GPT-family models perform strongest overall, but aggressive compaction still substantially reduces task accuracy. Failures can reflect distorted task state as well as omitted information--for example, treating unresolved issues as settled or forgetting established results, prompting incorrect or redundant actions.

### Source excerpt

arXiv:2611.25673v1 Announce Type: new Abstract: As language-model agents are applied to increasingly extended tasks, their working context continuously accumulates interaction history, tool observations, intermediate decisions, and task progress. Context compaction offers a practical way to reduce this growing history while allowing execution to continue, yet the compaction ability of modern LLMs remains insufficiently understood. We introduce CompAct-Bench, a benchmark for evaluating working-context compaction through its downstream effect on agent task completion. Starting from successful agent trajectories, we insert compaction boundaries at different stages of execution, compress the preceding working context with the model under evaluation, and let the same executor continue the original task. To enable fair comparison across compactors, we evaluate them under controlled retention ratios using a lightweight iterative ratio-adjustment procedure. CompAct-Bench covers information-seeking, software-engineering, and workspace-oriented tasks through BrowseComp, SWE-bench, and an internal CompanyBench evaluation. Our experiments show that working-context compaction remains highly lossy even for frontier models: Claude- and GPT-family models perform strongest overall, yet substantial degradation in downstream task accuracy persists under aggressive compaction. Further analysis reveals that failures are often caused not only by omitted information, but also by distorted task state, such as turning unresolved issues into apparent conclusions or forgetting previously established results, which can lead to incorrect decisions or redundant execution. These findings highlight working-context compaction as a distinct and still challenging ability for LLMs operating in extended agent tasks.

## MINTEval: Evaluating Memory under Multi-Target Interference in Long-Horizon Agent Systems

DevFeed: [MINTEval: Evaluating Memory under Multi-Target Interference in Long-Horizon Agent Systems](<https://devfeed.tech/articles/minteval-evaluating-memory-under-multi-target-interference-in-long-horizon-agent-systems-145334.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2605.18565>)

Author: Hyunji Lee, Justin Chih-Yao Chen, Joykirat Singh, Zaid Khan, Elias Stengel-Eskin, Mohit Bansal

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [Retrieval Augmented Generation (RAG)](<https://devfeed.tech/topics/retrieval-augmented-generation-rag.md>), [llm-reasoning](<https://devfeed.tech/topics/llm-reasoning.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agent-frameworks](<https://devfeed.tech/tags/agent-frameworks.md>), [aggregation](<https://devfeed.tech/tags/aggregation.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cl](<https://devfeed.tech/tags/cs-cl.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [frameworks](<https://devfeed.tech/tags/frameworks.md>), [github](<https://devfeed.tech/tags/github.md>), [horizon](<https://devfeed.tech/tags/horizon.md>), [llms](<https://devfeed.tech/tags/llms.md>), [long-context](<https://devfeed.tech/tags/long-context.md>), [long-context-llms](<https://devfeed.tech/tags/long-context-llms.md>), [memory](<https://devfeed.tech/tags/memory.md>), [reasoning](<https://devfeed.tech/tags/reasoning.md>), [recall](<https://devfeed.tech/tags/recall.md>), [retrieval](<https://devfeed.tech/tags/retrieval.md>), [revisions](<https://devfeed.tech/tags/revisions.md>), [robustness](<https://devfeed.tech/tags/robustness.md>), [state](<https://devfeed.tech/tags/state.md>), [tasks](<https://devfeed.tech/tags/tasks.md>)

### AI overview

MINTEval tests how seven systems--including long-context LLMs, RAG, and memory-augmented agents--handle evolving, interference-heavy information. Its 15.6k question-answer pairs span four domains and contexts averaging 138.8k tokens, up to 1.8M. Tasks test retrieval of a single target and aggregation across multiple pieces of evidence. Average accuracy was 27.9%, with multi-evidence reasoning especially difficult. The authors identify retrieval and memory construction as primary limitations; systems also struggled with earlier facts revised or disrupted by later context, and accuracy fell as intervening updates increased.

### Source excerpt

arXiv:2605.18565v3 Announce Type: replace-cross Abstract: Real-world agents operate over long and evolving horizons, where information is repeatedly updated and may interfere across memories, requiring accurate recall and aggregated reasoning over multiple pieces of information. However, existing benchmarks focus on static, independent recall and fail to capture these dynamic interactions between evolving memories. In this paper, we study how current memory-augmented agents perform in realistic, interference-heavy, long-horizon settings across diverse domains and question types. We introduce MINTEval (Long-Horizon Memory under INTerference Evaluation), a benchmark featuring (1) long, highly interconnected contexts with frequently updated information that induces substantial interference, (2) diverse domains (state tracking, multi-turn dialogue, Wikipedia revisions, and GitHub commits), enabling evaluation of domain generalization, and (3) diverse question types that assess robustness to interference, including (i) single-target recall tasks requiring retrieval of a specific target from long contexts, and (ii) multi-target aggregation tasks requiring reasoning over multiple relevant pieces of information. Overall, MINTEval has 15.6k question-answering pairs over long-horizon contexts averaging 138.8k tokens and extending up to 1.8M tokens per instance. We evaluate 7 representative systems, including vanilla long-context LLMs, RAG, and memory-augmented agent frameworks. Across all systems, we observe consistently low performance (avg. 27.9% accuracy), especially on questions requiring aggregated reasoning over multiple pieces of evidence. Our analysis shows that performance is primarily limited by retrieval and memory construction. Furthermore, current memory systems struggle to recall and reason over earlier facts that are revised or interfered with by subsequent context, with accuracy degrading as the number of intervening updates increases.

## AutoRAGBench: A Benchmark for LLM Agents Automating RAG System Development

DevFeed: [AutoRAGBench: A Benchmark for LLM Agents Automating RAG System Development](<https://devfeed.tech/articles/autoragbench-a-benchmark-for-llm-agents-automating-rag-system-development-145541.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.25582>)

Author: Hang Zhang, Yuzhi Pan, Ziyao Zhong, Chaokun Wang, Shuo Cao, Zeyu Huang, Yue Xue, Fang Niu, Bofan Xie, Guanchen Ge, Zecheng Li, Leqi Zheng, Ziyang Liu, Wei Wang, Miao Gu

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [Retrieval Augmented Generation (RAG)](<https://devfeed.tech/topics/retrieval-augmented-generation-rag.md>), [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [Large Language Model](<https://devfeed.tech/topics/llm.md>), [Benchmark](<https://devfeed.tech/topics/benchmark.md>)

Tags: [agents](<https://devfeed.tech/tags/agents.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cl](<https://devfeed.tech/tags/cs-cl.md>), [cs-ir](<https://devfeed.tech/tags/cs-ir.md>), [llm](<https://devfeed.tech/tags/llm.md>), [llm-agents](<https://devfeed.tech/tags/llm-agents.md>), [rag](<https://devfeed.tech/tags/rag.md>), [system-development](<https://devfeed.tech/tags/system-development.md>), [systems](<https://devfeed.tech/tags/systems.md>)

### AI overview

AutoRAGBench evaluates whether LLM agents can build and improve retrieval-augmented generation systems across nine development tasks. Agents worked in an open code space without starter code or prescribed strategies, with six hours and one H800 GPU per task; they analyzed data, implemented pipelines, and refined them using experimental feedback. Across nine configurations, GPT-6.1 Sol with Codex CLI had the highest average score (56.1), compared with 44.4 for the composite Human-SOTA baseline. Starting from human-designed systems on seven tasks, it improved scores by an average of 16.6 points. These results support the benchmark's finding that agents can automate effective RAG development under the tested conditions.

### Source excerpt

arXiv:2611.25582v1 Announce Type: cross Abstract: Retrieval-augmented generation (RAG) is a well-established paradigm for augmenting large language models (LLMs) with information retrieved from external corpora. However, developing effective RAG systems requires substantial engineering effort and iterative experimentation, as effective design choices vary across tasks and domains. With rapid advances in LLM agent capabilities, we ask how well these agents can autonomously build RAG systems and how effectively they can improve them through iteration. In this paper, we propose AutoRAGBench, the first benchmark to evaluate how well LLM agents develop RAG systems from scratch in an open code space. It comprises nine RAG system development tasks. For each task, agents build RAG systems from scratch without starter code or prescribed strategies, under a six-hour budget on a single H800 GPU. They analyze task data, implement RAG pipelines, and iteratively refine their designs based on experimental feedback. We evaluate nine agent configurations and analyze both the effectiveness of the resulting RAG systems and their development trajectories. GPT-6.1 Sol with Codex CLI achieves the highest average score of 56.1. The composite Human-SOTA baseline averages 44.4. These results show that LLM agents can automate effective RAG system development. When starting from a human-designed system on each of seven tasks, GPT-6.1 Sol with Codex CLI improves the score by 16.6 points on average.

## Cost-Aware Agentic Feature Optimization System for Fleet-Scale Ranking Models

DevFeed: [Cost-Aware Agentic Feature Optimization System for Fleet-Scale Ranking Models](<https://devfeed.tech/articles/cost-aware-agentic-feature-optimization-system-for-fleet-scale-ranking-models-145193.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.26531>)

Author: Chang Yang, Yao Xuan, Wendan Yan, Nhat Le, Jiang Liu, Longhao Jin, Zhen Peggy Yao, Alireza Vahdatpour, Djordje Gligorijevic

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>), [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Feature Engineering](<https://devfeed.tech/topics/feature-engineering.md>), [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [Combinatorial optimization](<https://devfeed.tech/topics/combinatorial-optimization.md>)

Tags: [agentic](<https://devfeed.tech/tags/agentic.md>), [asynchronous](<https://devfeed.tech/tags/asynchronous.md>), [blending](<https://devfeed.tech/tags/blending.md>), [budget](<https://devfeed.tech/tags/budget.md>), [candidate](<https://devfeed.tech/tags/candidate.md>), [cost](<https://devfeed.tech/tags/cost.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-ir](<https://devfeed.tech/tags/cs-ir.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [entropy](<https://devfeed.tech/tags/entropy.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [feature](<https://devfeed.tech/tags/feature.md>), [feature-selection](<https://devfeed.tech/tags/feature-selection.md>), [features](<https://devfeed.tech/tags/features.md>), [fleet](<https://devfeed.tech/tags/fleet.md>), [large](<https://devfeed.tech/tags/large.md>), [llm](<https://devfeed.tech/tags/llm.md>), [models](<https://devfeed.tech/tags/models.md>), [optimization](<https://devfeed.tech/tags/optimization.md>), [ranking](<https://devfeed.tech/tags/ranking.md>), [ranking-models](<https://devfeed.tech/tags/ranking-models.md>), [scale](<https://devfeed.tech/tags/scale.md>), [system](<https://devfeed.tech/tags/system.md>)

### AI overview

The system optimizes features across fleets of ranking models while respecting shared and per-model resource limits, including pipeline cost and serving throughput. LLM agents generate and shortlist cross features, while deterministic tools run and verify experiments. Several importance rankings are blended, using Pareto fronts for multi-task models; a Researcher agent then tunes selections through 10-25 asynchronous trials under resource caps, followed by trimming to meet pipeline budgets. Across production models, normalized entropy rose by 0.31% on average within each model's budget. Trimming reduced pipeline cost by 2.3% with no regression beyond measurement noise, and online A/B tests reported a 0.20% gain in a primary metric.

### Source excerpt

arXiv:2611.26531v1 Announce Type: cross Abstract: Large Web platforms operate large fleets of ranking models, each selecting features from a shared pool of tens of thousands. We present an end-to-end agentic system that optimizes features across the fleet under resource constraints shared across models and those set per model, such as feature pipeline cost and serving throughput (QPS), respectively. In this system, LLM agents propose hypotheses, and deterministic tools execute experiments and verify them. For each model, an LLM generates and shortlists cross features from a combinatorial candidate space, and all features are then ranked by several importance methods. Feature blending mixes these rankings into a single selection, using Pareto fronts across tasks for multi-task models. An LLM Researcher agent then tunes the feature selection in 10--25 asynchronous trials under resource caps, and a trimming step keeps the pipeline cost within budget. Across production models, the system improves normalized entropy by 0.31% on average within each model's budget, and the trimming step cuts pipeline cost by 2.3% with no regression beyond measurement noise. The optimized models were launched, and in online A/B tests they improved a primary online metric by 0.20%.

## ALIoTh: Orchestrating Multi-Agent LLMs for Autonomous IoT Penetration Testing

DevFeed: [ALIoTh: Orchestrating Multi-Agent LLMs for Autonomous IoT Penetration Testing](<https://devfeed.tech/articles/alioth-orchestrating-multi-agent-llms-for-autonomous-iot-penetration-testing-146047.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.25809>)

Author: Pironti Francesco Aurelio, Arena Luigi, Blefari Francesco, Lupinacci Matteo, Romeo Francesco, Maresca Alessio

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Cryptography and Security](<https://devfeed.tech/sources/arxiv-cryptography-and-security.md>)

Topics: [Hacking](<https://devfeed.tech/topics/hacking.md>), [Internet of things](<https://devfeed.tech/topics/iot.md>), [Multi Agent Systems](<https://devfeed.tech/topics/multi-agent-systems.md>), [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [Multi-step AI processes](<https://devfeed.tech/topics/multi-step-ai-processes.md>), [LLM security](<https://devfeed.tech/topics/llm-security.md>), [Enhancing AI model performance](<https://devfeed.tech/topics/enhancing-ai-model-performance.md>), [smartdevices](<https://devfeed.tech/topics/smartdevices.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [assessment](<https://devfeed.tech/tags/assessment.md>), [attack](<https://devfeed.tech/tags/attack.md>), [autonomous](<https://devfeed.tech/tags/autonomous.md>), [cameras](<https://devfeed.tech/tags/cameras.md>), [cs-cr](<https://devfeed.tech/tags/cs-cr.md>), [deployment](<https://devfeed.tech/tags/deployment.md>), [devices](<https://devfeed.tech/tags/devices.md>), [enumeration](<https://devfeed.tech/tags/enumeration.md>), [exploit](<https://devfeed.tech/tags/exploit.md>), [exploitation](<https://devfeed.tech/tags/exploitation.md>), [framework](<https://devfeed.tech/tags/framework.md>), [gathering](<https://devfeed.tech/tags/gathering.md>), [generation](<https://devfeed.tech/tags/generation.md>), [group-policy](<https://devfeed.tech/tags/group-policy.md>), [iiot](<https://devfeed.tech/tags/iiot.md>), [iot](<https://devfeed.tech/tags/iot.md>), [llms](<https://devfeed.tech/tags/llms.md>), [multi-agent](<https://devfeed.tech/tags/multi-agent.md>), [penetration-testing](<https://devfeed.tech/tags/penetration-testing.md>)

### AI overview

ALIoTh is a workflow-driven framework that coordinates specialized LLM agents through LangGraph and a Self-RAG architecture to automate IoT and IIoT penetration testing. Its described pipeline covers network, device, and service enumeration, vulnerability assessment, payload generation, reverse-shell deployment, and privilege escalation. Self-RAG gives agents contextual access to IoT security documentation, tool manuals, and known attack techniques, supporting scenarios involving multiple devices and group-policy manipulation. The supplied text describes the approach but provides no evaluation results or comparative performance.

### Source excerpt

arXiv:2611.25809v1 Announce Type: new Abstract: Penetration testing in Internet of Things (IoT) environments, especially those composed of various different devices, e.g., smart plugs, smart cameras, and smart doors, presents great challenges across the phases of information gathering, attack surface enumeration, privilege escalation, and lateral movement. This paper introduces ALIoTh, a multi-agent, workflow-driven penetration testing framework that integrates Large Language Models agents within a Self-Retrieval-Augmented Generation (Self-RAG) architecture to autonomously identify and exploit vulnerabilities in IoT and Industrial-IoT (IIoT) environments. By leveraging state-of-the-art language models and orchestrating specialized agents via the LangGraph framework, ALIoTh dynamically plans and executes a full-spectrum offensive pipeline. This includes network, devices, and service enumeration, vulnerability assessment, payload generation, reverse shell deployment, and privilege escalation on each device. The Self-RAG mechanism equips expert agents with contextual access to IoT security knowledge, e.g., related documentation, tool manuals, and known attack techniques, enhancing their reasoning capabilities in scenarios involving multiple device exploitation and group policy manipulation.

## MKA: Evaluating Multimodal Knowledge Agents on end-to-end professional tasks

DevFeed: [MKA: Evaluating Multimodal Knowledge Agents on end-to-end professional tasks](<https://devfeed.tech/articles/mka-evaluating-multimodal-knowledge-agents-on-end-to-end-professional-tasks-145486.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.25228>)

Author: Anima Singh, Suhas Yogin, Weicheng Kuo, Ivor Rendulic, Ivan Jurin, Antoine Yang, Oran Kelly, Rohan Doshi, Hammad Ayyubi, Ioana Bica, Sergi Caelles, Matthew Conniff, Dan Graur, Anastasija Ili\\'c, Ahmet

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [Frontier AI](<https://devfeed.tech/topics/frontier-ai.md>), [Complex AI tasks](<https://devfeed.tech/topics/complex-ai-tasks.md>), [gaia](<https://devfeed.tech/topics/gaia.md>), [gpt-6-astra](<https://devfeed.tech/topics/gpt-6-astra.md>), [Claude](<https://devfeed.tech/topics/claude.md>)

Tags: [benchmark](<https://devfeed.tech/tags/benchmark.md>), [blueprints](<https://devfeed.tech/tags/blueprints.md>), [charts](<https://devfeed.tech/tags/charts.md>), [claude-fable-5-2](<https://devfeed.tech/tags/claude-fable-5-2.md>), [claude-opus-5-5](<https://devfeed.tech/tags/claude-opus-5-5.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cv](<https://devfeed.tech/tags/cs-cv.md>), [files](<https://devfeed.tech/tags/files.md>), [format](<https://devfeed.tech/tags/format.md>), [frontier-models](<https://devfeed.tech/tags/frontier-models.md>), [gemini-4-carbon](<https://devfeed.tech/tags/gemini-4-carbon.md>), [gpt-6-astra](<https://devfeed.tech/tags/gpt-6-astra.md>), [knowledge-agents](<https://devfeed.tech/tags/knowledge-agents.md>), [knowledge-work](<https://devfeed.tech/tags/knowledge-work.md>), [maps](<https://devfeed.tech/tags/maps.md>), [multimodal](<https://devfeed.tech/tags/multimodal.md>)

### AI overview

MKA evaluates whether autonomous agents can combine information across visually dense files and produce professional deliverables. Its 200 tasks span 10 economic sectors and formats including spreadsheets, presentations, PDFs, documents, HTML, and images. An agentic autorater applies expert-weighted rubrics; in an expert-graded study, its criterion judgments agreed with domain experts 96.2% of the time (Cohen's κ = 0.91). Among four evaluated frontier models, Claude Fable 5.2 scored highest on Climb-WSR (56.6%) and overall WSR (73.3%), followed by Gemini 4 Carbon (54.3% and 71.9%); Carbon used 5.7x fewer tokens. The authors note that even the strongest models remain below 57% Climb-WSR, indicating substantial remaining headroom. The supplied evidence is an abstract, so it does not establish full-paper details.

### Source excerpt

arXiv:2611.25228v1 Announce Type: new Abstract: We introduce the Multimodal Knowledge Agent (MKA) benchmark, which measures how well autonomous agents perform end-to-end professional knowledge work that depends on visual understanding. Each task requires an agent to extract and combine information from multiple visually dense files, such as CAD blueprints, electronics schematics, radiographs, complex charts and geospatial maps, and to produce a standalone deliverable in .xlsx, .pptx, .pdf, .docx, .html or .png format. Existing benchmarks leave this capability largely untested: removing image-viewing tools leaves Gemini 4 Carbon's score on GDPval unchanged, and single-format evaluations such as GDP.pdf and GDP.xlsx do not require reasoning across files or modalities. MKA contains 200 tasks across 10 economic sectors. Deliverables are graded against expert-weighted rubrics by a sandboxed agentic autorater that agrees with human domain experts on 96.2% of criterion judgments in our expert-graded study (Cohen's $\kappa = 0.91$), scored by weighted satisfaction rate (WSR) overall and on frontier criteria unsaturated by prior-generation models (Climb-WSR). Across the frontier models we evaluate (Gemini 4 Carbon, Claude Fable 5.2, Claude Opus 5.5, and GPT-6 Astra), Claude Fable 5.2 (56.6% Climb-WSR, 73.3% Overall WSR) and Gemini 4 Carbon (54.3% Climb-WSR, 71.9% Overall WSR) achieve the highest scores, while Carbon uses $5.7\times$ fewer tokens. Even for the strongest models, Climb-WSR remains below 57%, leaving substantial headroom for multimodal agents.

## TwinLab-Bench: Evaluating AI Scientists in Agentic Twins of Real Labs

DevFeed: [TwinLab-Bench: Evaluating AI Scientists in Agentic Twins of Real Labs](<https://devfeed.tech/articles/twinlab-bench-evaluating-ai-scientists-in-agentic-twins-of-real-labs-145156.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.26248>)

Author: Yang Wan, Linchao Zhu

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>), [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [AI Chat](<https://devfeed.tech/topics/ai-chat.md>), [Claude](<https://devfeed.tech/topics/claude.md>)

Tags: [agentic](<https://devfeed.tech/tags/agentic.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [bench](<https://devfeed.tech/tags/bench.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cl](<https://devfeed.tech/tags/cs-cl.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [environment](<https://devfeed.tech/tags/environment.md>), [experiment](<https://devfeed.tech/tags/experiment.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [frontier](<https://devfeed.tech/tags/frontier.md>), [frontier-models](<https://devfeed.tech/tags/frontier-models.md>), [grounded](<https://devfeed.tech/tags/grounded.md>), [laboratory](<https://devfeed.tech/tags/laboratory.md>), [labs](<https://devfeed.tech/tags/labs.md>), [language-model](<https://devfeed.tech/tags/language-model.md>), [life-sciences](<https://devfeed.tech/tags/life-sciences.md>), [measurements](<https://devfeed.tech/tags/measurements.md>), [memory](<https://devfeed.tech/tags/memory.md>), [model](<https://devfeed.tech/tags/model.md>), [models](<https://devfeed.tech/tags/models.md>)

### AI overview

TwinLab-Bench tests whether AI agents can design experiments and obtain readings grounded in real laboratory records, extending beyond benchmarks that assess fixed workflows or choices among predefined measurements. It covers 52 tasks across physical sciences, life sciences, and engineering, with more than 100 experimental action types. For each task, a TwinLab agent executes the tested agent's designed experiments using laboratory materials and record-based calculators; reports are scored against hidden answers from real lab records. The authors say this setup nearly eliminates readings that deviate from the records and inconsistent responses to repeated requests compared with a language model given only the task description; closed-book probes address answers recalled from memory. The tasks remain difficult for frontier models: Claude Fable 5.2, the strongest reported, averaged 45.0. The excerpt does not specify the score scale.

### Source excerpt

arXiv:2611.26248v1 Announce Type: cross Abstract: AI agents are increasingly expected to act as AI scientists, which requires three increasingly open abilities: carrying out a given computational workflow correctly, choosing the next measurements from a given design space, and designing their own experiments. Existing benchmarks mainly evaluate the first two abilities, in a terminal or in a closed loop that selects from predefined candidates. Evaluating the third ability requires reliably executing an experiment that an agent designs, with readings grounded in a real laboratory. We introduce TwinLab-Bench, which evaluates AI scientists on self-designed experiments in 52 tasks from laboratories in the physical sciences, the life sciences, and engineering. The benchmark supports more than 100 experimental action types, and the agent's report is scored by comparison with a hidden answer taken from real laboratory records. Each task has its own TwinLab agent, an agent that serves as the environment and executes the experiments that the tested agent designs, reading the raw materials of the laboratory and calling calculators that are built from its records. Compared with a language model that sees only the task description, the TwinLab agent almost eliminates readings that deviate from the real records and inconsistent answers to a repeated request, and closed-book probes control for answers recalled from memory. We find that these tasks remain hard for frontier models: the strongest, Claude Fable 5.2, averages a score of only 45.0, and we conduct an empirical study of where it fails.

## Toward a Science of AI Agent Societies

DevFeed: [Toward a Science of AI Agent Societies](<https://devfeed.tech/articles/toward-a-science-of-ai-agent-societies-145508.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.25404>)

Author: Geon Lee, Fanchen Bu, Soo Yong Lee, Sunwoo Kim, Kijung Shin

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [Multi Agent Systems](<https://devfeed.tech/topics/multi-agent-systems.md>), [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [future of AI](<https://devfeed.tech/topics/future-of-ai.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [autonomy](<https://devfeed.tech/tags/autonomy.md>), [behaviors](<https://devfeed.tech/tags/behaviors.md>), [challenges](<https://devfeed.tech/tags/challenges.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [governance](<https://devfeed.tech/tags/governance.md>), [implementation](<https://devfeed.tech/tags/implementation.md>), [multi-agent-systems](<https://devfeed.tech/tags/multi-agent-systems.md>), [objectives](<https://devfeed.tech/tags/objectives.md>), [research](<https://devfeed.tech/tags/research.md>), [scale](<https://devfeed.tech/tags/scale.md>), [society](<https://devfeed.tech/tags/society.md>)

### AI overview

The paper proposes studying AI agent societies as a distinct research object, rather than simply scaling up individual agents or simulating human society. It defines four core properties for such societies: individualized objectives, rules and governance, autonomy, and scale and complexity. The authors identify economic behavior, conflict of interest, unsafe or unethical behavior, and system-level behavior as areas to investigate. They highlight property parameterization, parameter balancing, and robust implementation as technical challenges, and assess existing multi-agent systems as satisfying only subsets of the proposed properties. The abstract says progress could support useful models and inform platform design, evaluation, and governance.

### Source excerpt

arXiv:2611.25404v1 Announce Type: new Abstract: AI agents are rapidly evolving from isolated personal assistants into networked actors that interact with one another at scale. We envision the emergence of AI agent societies, with their own social and economic dynamics, as a new research frontier. We argue that AI agent societies should be studied as a distinct object of inquiry: neither simply larger collections of individual agents nor merely simulations of human society. To formalize this perspective, we propose four core properties that a valid AI agent society should satisfy: individualized objectives, rules and governance, autonomy, and scale and complexity. Building on this framework, we identify four classes of societal behaviors worth studying in AI agent societies: economic behaviors, behaviors under conflict-of-interest, unsafe and unethical behaviors, and system-level behaviors. We then outline key technical challenges--including property parameterization, parameter balancing, and robust implementation--and argue that progress on these challenges could enable scientifically informative and practically useful models of AI agent societies. Finally, we revisit existing multi-agent systems through the lens of the proposed core properties, show that they instantiate only subsets of them, and discuss implications for platform design, evaluation, and governance.

## NariQ: Sub-3-Bit Quantization for Long-Horizon Agentic Tasks

DevFeed: [NariQ: Sub-3-Bit Quantization for Long-Horizon Agentic Tasks](<https://devfeed.tech/articles/nariq-sub-3-bit-quantization-for-long-horizon-agentic-tasks-145146.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.26157>)

Author: Doyeob Kim

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [quantization](<https://devfeed.tech/topics/quantization.md>), [Post-training optimization](<https://devfeed.tech/topics/post-training-optimization.md>), [Large Language Model](<https://devfeed.tech/topics/llm.md>), [AI Agent](<https://devfeed.tech/topics/ai-agent.md>)

Tags: [agentic](<https://devfeed.tech/tags/agentic.md>), [agentic-benchmark](<https://devfeed.tech/tags/agentic-benchmark.md>), [agentic-tasks](<https://devfeed.tech/tags/agentic-tasks.md>), [aime](<https://devfeed.tech/tags/aime.md>), [bf16](<https://devfeed.tech/tags/bf16.md>), [bits](<https://devfeed.tech/tags/bits.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [deepswe](<https://devfeed.tech/tags/deepswe.md>), [gguf](<https://devfeed.tech/tags/gguf.md>), [inference](<https://devfeed.tech/tags/inference.md>), [post-training](<https://devfeed.tech/tags/post-training.md>), [quantization](<https://devfeed.tech/tags/quantization.md>), [qwen3](<https://devfeed.tech/tags/qwen3.md>), [qwen3-8-27b](<https://devfeed.tech/tags/qwen3-8-27b.md>), [terminal-bench](<https://devfeed.tech/tags/terminal-bench.md>), [trellis](<https://devfeed.tech/tags/trellis.md>)

### AI overview

NariQ targets a gap in low-bit quantization for long-horizon agent tasks: methods calibrated on short web text can retain performance on AIME while doing substantially worse on DeepSWE. Its trellis-coded post-training method calibrates on tokens generated by the BF16 model during agent sessions. For Qwen3.8-27B, the 3.0- and 2.5-bit-per-weight versions are reported to match BF16 on DeepSWE and Terminal Bench, as well as AIME and RealWorldQA; the 2.0-bit checkpoint trails that performance. The authors describe these as the first quantizations at or below 3 bits shown to match full precision on long-horizon agent tasks. They provide GGUF files and inference through a llama.cpp fork.

### Source excerpt

arXiv:2611.26157v1 Announce Type: new Abstract: Long-horizon agentic tasks require a model to call tools over many steps at contexts of 100K tokens or more, yet many low-bit quantization methods are calibrated on short web text and judged on perplexity. This mismatch is costly: existing quantizations of Qwen3.8-27B at or below 3 bits per weight match BF16 on shorter tasks like AIME, yet in published results solve at most 72% as many tasks as BF16 on DeepSWE, a long-horizon agentic benchmark. We propose NariQ, a trellis-coded post-training quantization method that calibrates on tokens generated by the BF16 model in agent sessions. On Qwen3.8-27B, NariQ at 3.0 and 2.5 bits per weight stays level with BF16 on two long-horizon agentic benchmarks, DeepSWE and Terminal Bench, as well as on AIME and RealWorldQA. Our checkpoint at 2.0 bits per weight trails this performance. To our knowledge, these are the first quantizations at or below 3 bits per weight shown to match their full-precision version on long-horizon agentic tasks. We release the files as Qwen3.8-27B-NariQ-GGUF (https://huggingface.co/nari-labs/Qwen3.8-27B-NariQ-GGUF), with inference available through our llama.cpp fork at https://github.com/nari-labs/nari-llama.cpp.

## Duopra: From Full-Duplex Conversation to Full-Duplex Agency

DevFeed: [Duopra: From Full-Duplex Conversation to Full-Duplex Agency](<https://devfeed.tech/articles/duopra-from-full-duplex-conversation-to-full-duplex-agency-145586.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.25773>)

Author: Chiyu Chen, Pengzhou Cheng, Yuezhao Ma, Xiaocheng Li, Can Liu, Haonan Hu, Yao Yao, Yifei Yang, Yijie Lu, Taolei Wang, Yansi Li, Muhuan Wu, Chenglu Sun, Gongshen Liu, Zhuosheng Zhang

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [function calling](<https://devfeed.tech/topics/function-calling.md>), [post-training](<https://devfeed.tech/topics/post-training.md>), [Multi-step AI processes](<https://devfeed.tech/topics/multi-step-ai-processes.md>)

Tags: [audio](<https://devfeed.tech/tags/audio.md>), [backend](<https://devfeed.tech/tags/backend.md>), [bench](<https://devfeed.tech/tags/bench.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [execution](<https://devfeed.tech/tags/execution.md>), [frontend](<https://devfeed.tech/tags/frontend.md>), [full-duplex-bench](<https://devfeed.tech/tags/full-duplex-bench.md>), [interaction](<https://devfeed.tech/tags/interaction.md>), [interface](<https://devfeed.tech/tags/interface.md>), [latency](<https://devfeed.tech/tags/latency.md>), [low-latency](<https://devfeed.tech/tags/low-latency.md>), [model](<https://devfeed.tech/tags/model.md>), [natural-language](<https://devfeed.tech/tags/natural-language.md>), [pass](<https://devfeed.tech/tags/pass.md>), [pipeline](<https://devfeed.tech/tags/pipeline.md>), [protocol](<https://devfeed.tech/tags/protocol.md>), [realtime](<https://devfeed.tech/tags/realtime.md>), [repair](<https://devfeed.tech/tags/repair.md>), [request](<https://devfeed.tech/tags/request.md>), [requires](<https://devfeed.tech/tags/requires.md>), [v3](<https://devfeed.tech/tags/v3.md>)

### AI overview

Duopra addresses a gap in full-duplex agents: fluent turn-taking does not ensure timely, complete tool requests or corrections. It treats the backend as a tool that receives a self-contained natural-language request, using one action interface for direct calls and backend delegation. Training combines protocol SFT, full-duplex reinforcement learning for timely and complete execution, and projector-only repair of talker conditioning after thinker updates. On FDB v3, the authors report Pass@1 rising from 4.7% to 44.7% with the same backend; Duopra reaches 47.7% with direct tool calls (26.0% after SFT) and 49.3% with a DeepSeek backend. It also improves three FDB v1.0 takeover rates over the base model. The excerpt supports the authors' claim that task agency can coexist with key conversational agency.

### Source excerpt

arXiv:2611.25773v1 Announce Type: new Abstract: Full-duplex interaction models exhibit strong conversational agency, enabling natural, low-latency turn-taking, backchannels, and interruptions. However, completing tasks further requires task agency: timely tool invocation and argument revision following user corrections. Frontend-backend agents separate initiation from execution: the frontend decides whether and what to request, while the backend determines how. Execution can be offloaded, but unsent requests cannot be executed, and incomplete ones can misdirect the backend. On Full-Duplex-Bench v3 (FDB v3), most failures of Realtime-Venus-Audio + Harness in our streaming pipeline occur before a domain tool executes. We introduce Duopra, a full-duplex interaction model that learns task execution alongside conversational control. Duopra treats the backend as a tool, accepting a self-contained natural-language request. Direct calls and backend delegation share an action interface and requirements for faithful initiation and correct execution. Its three-stage training combines protocol SFT, full-duplex RL rewarding timely and complete execution, and projector-only repair of talker conditioning after thinker updates. With the same backend, Pass@1 rises from 4.7% to 44.7%; Duopra achieves 47.7% calling tools directly (26.0% after SFT) and 49.3% with a DeepSeek backend. It improves three FDB v1.0 takeover rates over the base model. These results show that task agency can coexist with key conversational agency.

## Beyond Weight Updates: Evolving Agent Harnesses for Domain Specialization

DevFeed: [Beyond Weight Updates: Evolving Agent Harnesses for Domain Specialization](<https://devfeed.tech/articles/beyond-weight-updates-evolving-agent-harnesses-for-domain-specialization-145504.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.25380>)

Author: BoHao Chen

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [Agent Harness](<https://devfeed.tech/topics/agent-harness.md>), [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [llm-reasoning](<https://devfeed.tech/topics/llm-reasoning.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>)

Tags: [adapters](<https://devfeed.tech/tags/adapters.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agent-harness](<https://devfeed.tech/tags/agent-harness.md>), [agent-harnesses](<https://devfeed.tech/tags/agent-harnesses.md>), [coding-agent](<https://devfeed.tech/tags/coding-agent.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [domain](<https://devfeed.tech/tags/domain.md>), [gepa](<https://devfeed.tech/tags/gepa.md>), [gpu](<https://devfeed.tech/tags/gpu.md>), [llm](<https://devfeed.tech/tags/llm.md>), [llm-agents](<https://devfeed.tech/tags/llm-agents.md>), [planning](<https://devfeed.tech/tags/planning.md>), [policy](<https://devfeed.tech/tags/policy.md>), [prompt](<https://devfeed.tech/tags/prompt.md>), [reasoning](<https://devfeed.tech/tags/reasoning.md>), [reinforcement-learning](<https://devfeed.tech/tags/reinforcement-learning.md>), [specialization](<https://devfeed.tech/tags/specialization.md>), [swe-bench](<https://devfeed.tech/tags/swe-bench.md>), [terminal-bench](<https://devfeed.tech/tags/terminal-bench.md>), [weight](<https://devfeed.tech/tags/weight.md>)

### AI overview

The study evolves an agent's system prompt, tool schemas, skills, and planning scaffolding while keeping its policy weights frozen. It frames this as an RL-style outer loop and uses GEPA, a reflective evolutionary optimizer, to produce domain-specific harness adapters that require no GPU training and can be combined with weight-based RL. In experiments with Codev on SWE-bench Verified, OfficeBench, and Terminal-Bench, weight updates gave the strongest gains when used alone; harness evolution recovered most of those gains at a small fraction of the compute, and combining the approaches performed best. The authors also analyze SFT specialization's shortened reasoning chains and increased truncation, arguing that harness adapters avoid this failure mode by construction. These findings are limited to the reported agent and benchmarks.

### Source excerpt

arXiv:2611.25380v1 Announce Type: new Abstract: Domain specialization of LLM agents is usually framed as a weight-update problem: collect trajectories, run reinforcement learning (RL) or supervised fine-tuning (SFT), and deploy a new checkpoint. We study a complementary axis: evolving the \textbf{agent harness}---the system prompt, tool schemas, skill library, and planning scaffolding---while keeping the policy frozen. We cast harness evolution as an RL-style outer loop and instantiate it with GEPA, a reflective evolutionary optimizer. The resulting harness adapters require no GPU training, are loadable per vertical domain, and compose with weight-based RL. Experiments with Codev, a Claude Code-style coding agent, on SWE-bench Verified, OfficeBench, and Terminal-Bench show a consistent pattern: weight updates give the strongest single-path gains, harness evolution recovers most of those gains at a small fraction of the compute, and combining both paths is best overall. We also analyze a failure mode of SFT-based specialization---shortened reasoning chains and increased truncation---and argue that harness adapters avoid it by construction. Our results position harness evolution as a practical, non-destructive complement to weight updates for vertical agents.

## Why agent governance starts with observability

DevFeed: [Why agent governance starts with observability](<https://devfeed.tech/articles/why-agent-governance-starts-with-observability-146227.md>)

Original publisher: [Read original article](<https://www.langchain.com/blog/agent-governance-observability>)

Author: LangChain Accounts

Published: 2026-11-20T00:57:43Z

Content type: article

Language: en

Sources: [LangChain Blog](<https://devfeed.tech/sources/langchain-blog.md>)

Topics: [agent observability](<https://devfeed.tech/topics/agent-observability.md>), [LLM observability](<https://devfeed.tech/topics/llm-observability.md>), [ai-governance](<https://devfeed.tech/topics/ai-governance.md>), [AI Agent](<https://devfeed.tech/topics/ai-agent.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agent-observability](<https://devfeed.tech/tags/agent-observability.md>), [agent-security](<https://devfeed.tech/tags/agent-security.md>), [ai](<https://devfeed.tech/tags/ai.md>), [governance](<https://devfeed.tech/tags/governance.md>), [observability](<https://devfeed.tech/tags/observability.md>)

### AI overview

Agent governance depends on observability: centralized traces and run trajectories let teams see which agents are operating, inspect their actions, and check whether guardrails hold. The excerpt distinguishes preventative controls--runtime limits and policies, plus pre-deployment evaluations and red-teaming--from detective measures such as online evaluations, dashboards, alerts, drift checks, and human review. Traces support both: they help explain failures before release and audit production behavior, including alternate paths that may bypass intended restrictions. Production findings can feed back into tighter guardrails or regression tests. The article also names LangSmith Insights reports for clustering trace patterns and LangSmith Engine for analyzing recurring failures and testing proposed fixes. These are presented as vendor capabilities; the excerpt does not establish comparative performance against other tools.

### Source excerpt

A strong AI agent governance strategy

## Apple may be developing its own AI agent that runs your Mac

DevFeed: [Apple may be developing its own AI agent that runs your Mac](<https://devfeed.tech/articles/apple-may-be-developing-its-own-ai-agent-that-runs-your-mac-144606.md>)

Original publisher: [Read original article](<https://techcrunch.com/2026/11/19/apple-may-be-developing-its-own-ai-agent-that-runs-your-mac/>)

Author: Lucas Ropek

Published: 2026-11-20T00:56:01Z

Content type: news

Language: en

Sources: [TechCrunch | AI News & Artificial Intelligence](<https://devfeed.tech/sources/techcrunch-ai-news-artificial-intelligence.md>)

Topics: [AI Agent](<https://devfeed.tech/topics/ai-agent.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [anthropic](<https://devfeed.tech/tags/anthropic.md>), [apis](<https://devfeed.tech/tags/apis.md>), [app](<https://devfeed.tech/tags/app.md>), [apple](<https://devfeed.tech/tags/apple.md>), [apps](<https://devfeed.tech/tags/apps.md>), [bloomberg](<https://devfeed.tech/tags/bloomberg.md>), [browsers](<https://devfeed.tech/tags/browsers.md>), [code](<https://devfeed.tech/tags/code.md>), [coding](<https://devfeed.tech/tags/coding.md>), [command](<https://devfeed.tech/tags/command.md>), [commands](<https://devfeed.tech/tags/commands.md>), [iphone](<https://devfeed.tech/tags/iphone.md>), [meta](<https://devfeed.tech/tags/meta.md>), [meta-muse](<https://devfeed.tech/tags/meta-muse.md>), [muse](<https://devfeed.tech/tags/muse.md>), [openai](<https://devfeed.tech/tags/openai.md>)

### AI overview

Apple appears to be developing an AI agent that can use a computer across apps, browsers, command lines, code, and APIs, handling multi-step work and choosing tools along the way. The report is based on a job listing for a senior engineering manager; Apple did not respond to TechCrunch's request for comment. Bloomberg says the project remains in early stages, though Apple has eyed a launch as soon as the end of next year. At the same time, Apple plans additional controls in macOS 27.3 to discourage granting third-party agents full access, requiring explicit user action for that access.

### Source excerpt

The iPhone maker appears to be building agents that can compete with the likes of Meta Muse, Instinct, Claude, and OpenAI Dots.

## Who's accountable when the worker is an agent?

DevFeed: [Who's accountable when the worker is an agent?](<https://devfeed.tech/articles/who-s-accountable-when-the-worker-is-an-agent-146879.md>)

Original publisher: [Read original article](<https://webflow.sysdig.com/blog/whos-accountable-when-the-worker-is-an-agent>)

Author: Matt Adams

Published: 2026-11-20T00:00:00Z

Content type: article

Language: en

Sources: [Sysdig](<https://devfeed.tech/sources/sysdig-blog.md>)

Topics: [ai-governance](<https://devfeed.tech/topics/ai-governance.md>), [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [shadow AI](<https://devfeed.tech/topics/shadow-ai.md>), [Access Control](<https://devfeed.tech/topics/access-control.md>)

Tags: [accountability](<https://devfeed.tech/tags/accountability.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [ai-security](<https://devfeed.tech/tags/ai-security.md>), [ai-tools](<https://devfeed.tech/tags/ai-tools.md>), [ai-workload-security](<https://devfeed.tech/tags/ai-workload-security.md>), [controls](<https://devfeed.tech/tags/controls.md>), [ground-truth-london](<https://devfeed.tech/tags/ground-truth-london.md>)

### AI overview

AI agents do not bear responsibility for consequences, so accountability must remain tied to the person who set them running and to evidence the agent cannot alter. The article recommends first inventorying agents, data flows, and the people behind them, including informal prompt or skill files. Enforce limits outside the agent and its harness: isolate it where practical, grant only task-required access, observe actual system activity, and retain the ability to stop it and roll back. The agent's own account can show its intent but should not be the sole record. These controls can support broader approval--for example, a coding agent might receive a read-only production database role and have every query logged. The author places approval with the data/AI function or adoption-program owner, with security advising and owning guardrails rather than making the final decision. The excerpt says no complete accountability model exists yet.

### Source excerpt

If something goes wrong with your work, you know you're accountable. But what takes the place of accountability when an AI agent is doing the work?

## Outage, Attack, or AI Agent Drift: Lessons from the Microsoft 365 Outage

DevFeed: [Outage, Attack, or AI Agent Drift: Lessons from the Microsoft 365 Outage](<https://devfeed.tech/articles/outage-attack-or-ai-agent-drift-lessons-from-the-microsoft-365-outage-145970.md>)

Original publisher: [Read original article](<https://www.veeam.com/blog/microsoft-365-outage-2026-lessons.html>)

Author: Edward Watson

Published: 2026-11-19T22:07:36Z

Content type: opinion

Language: en

Sources: [Veeam Software Official Blog](<https://devfeed.tech/sources/veeam-software-official-blog.md>)

Topics: [microsoft 365](<https://devfeed.tech/topics/microsoft-365.md>), [Resilience](<https://devfeed.tech/topics/resilience.md>), [Cybersecurity](<https://devfeed.tech/topics/cybersecurity.md>), [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [ai security](<https://devfeed.tech/topics/ai-security.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [attack](<https://devfeed.tech/tags/attack.md>), [business](<https://devfeed.tech/tags/business.md>), [copilot](<https://devfeed.tech/tags/copilot.md>), [data-resilience](<https://devfeed.tech/tags/data-resilience.md>), [defender](<https://devfeed.tech/tags/defender.md>), [disaster-recovery](<https://devfeed.tech/tags/disaster-recovery.md>), [exchange](<https://devfeed.tech/tags/exchange.md>), [exchange-online](<https://devfeed.tech/tags/exchange-online.md>), [incident](<https://devfeed.tech/tags/incident.md>), [microsoft](<https://devfeed.tech/tags/microsoft.md>), [microsoft-365](<https://devfeed.tech/tags/microsoft-365.md>), [microsoft-365-outage](<https://devfeed.tech/tags/microsoft-365-outage.md>), [outage](<https://devfeed.tech/tags/outage.md>), [ransomware](<https://devfeed.tech/tags/ransomware.md>), [saas-backup](<https://devfeed.tech/tags/saas-backup.md>), [sharepoint](<https://devfeed.tech/tags/sharepoint.md>), [teams](<https://devfeed.tech/tags/teams.md>), [veeam](<https://devfeed.tech/tags/veeam.md>)

### AI overview

Microsoft's roughly 24-hour Microsoft 365 outage left data intact but inaccessible, exposing how dependent business continuity can be on shared cloud services. The article argues for two recovery goals: keep critical work and communication going while a platform is down, then restore data to a verified, trusted state. It recommends mapping SaaS dependencies, keeping independent copies and out-of-band communications, isolating immutable backups, checking recovery points before restoration, and rehearsing both recovery timelines with business leaders. It also distinguishes infrastructure failure, cyberattacks, and AI-agent mistakes: agent-driven changes may require identifying affected data and selectively reversing changes instead of rolling back an entire system. Veeam says its Microsoft 365 offerings support access to backed-up data during outages and selective undo of supported agent-modified files; these require advance setup or available restore points. The excerpt notes these tools do not replace identity controls, permission governance, or continuity planning.

### Source excerpt

On Thursday, Nov. 12, Microsoft 365 went dark for much of the world. Problems started with Exchange Online around 11:20 a.m. ET, then spread to Teams, SharePoint, Copilot, and even Defender XDR. Outage reports came in from 74 countries and territories, and, by Microsoft's own incident record, the disruption ran just shy of 24 hours. The post Outage, Attack, or AI Agent Drift: Lessons from the Microsoft 365 Outage appeared first on Veeam Software Official Blog.

## Your AI Agent Is Only Seeing Half The Answer

DevFeed: [Your AI Agent Is Only Seeing Half The Answer](<https://devfeed.tech/articles/your-ai-agent-is-only-seeing-half-the-answer-145830.md>)

Original publisher: [Read original article](<https://www.starburst.io/blog/your-ai-agent-is-only-seeing-half-the-answer>)

Published: 2026-11-19T18:00:05Z

Content type: article

Language: en

Sources: [Starburst Data](<https://devfeed.tech/sources/starburst-data.md>)

Topics: [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [AI Chat](<https://devfeed.tech/topics/ai-chat.md>), [AI Context Management](<https://devfeed.tech/topics/ai-context-management.md>), [Apache Iceberg](<https://devfeed.tech/topics/apache-iceberg.md>)

Tags: [access-control](<https://devfeed.tech/tags/access-control.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [ai-data](<https://devfeed.tech/tags/ai-data.md>), [ai-data-engine](<https://devfeed.tech/tags/ai-data-engine.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [api](<https://devfeed.tech/tags/api.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [auditing](<https://devfeed.tech/tags/auditing.md>), [business](<https://devfeed.tech/tags/business.md>), [compliance](<https://devfeed.tech/tags/compliance.md>), [connectors](<https://devfeed.tech/tags/connectors.md>), [console](<https://devfeed.tech/tags/console.md>), [content](<https://devfeed.tech/tags/content.md>), [context](<https://devfeed.tech/tags/context.md>), [contracts](<https://devfeed.tech/tags/contracts.md>), [control](<https://devfeed.tech/tags/control.md>)

### AI overview

Starburst AIDA and NetApp AIDE are presented as a way to answer questions that require both database values and document context, such as matching contract renewal dates and spend with notice clauses. AIDA plans the question, sends structured queries through Starburst's federated SQL engine and document searches through AIDE's MCP server, then joins the results and cites source versions. Starburst is described as querying structured and semi-structured data in place across 50+ connectors, with row- and column-level controls, masking, and auditing. AIDE makes files on NetApp storage searchable, preserves file permissions, and exposes content through MCP; its Metadata Services catalog reads metadata without copying files or using GPUs. The stated tradeoff is that this approach avoids a separate external index, but its document capabilities depend on files being accessible through AIDE and NetApp storage interfaces. The excerpt recommends starting with one cross-source question and a joint architecture workshop.

### Source excerpt

Most teams building AI agents hit the same problem. Everything isn't in the same place. ...

## Meta's AI agent Muse is mixing up conversations and exposing private data, report says

DevFeed: [Meta's AI agent Muse is mixing up conversations and exposing private data, report says](<https://devfeed.tech/articles/meta-s-ai-agent-muse-is-mixing-up-conversations-and-exposing-private-data-report-says-143920.md>)

Original publisher: [Read original article](<https://techcrunch.com/2026/11/19/metas-ai-agent-muse-is-mixing-up-conversations-and-exposing-private-data-report-says/>)

Author: Aisha Malik

Published: 2026-11-19T17:38:40Z

Content type: news

Language: en

Sources: [TechCrunch | AI News & Artificial Intelligence](<https://devfeed.tech/sources/techcrunch-ai-news-artificial-intelligence.md>)

Topics: [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [Conversational AI](<https://devfeed.tech/topics/conversational-ai.md>), [sensitive data](<https://devfeed.tech/topics/sensitive-data.md>)

Tags: [account](<https://devfeed.tech/tags/account.md>), [advice](<https://devfeed.tech/tags/advice.md>), [agent](<https://devfeed.tech/tags/agent.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [app](<https://devfeed.tech/tags/app.md>), [apps](<https://devfeed.tech/tags/apps.md>), [bank](<https://devfeed.tech/tags/bank.md>), [career](<https://devfeed.tech/tags/career.md>), [career-advice](<https://devfeed.tech/tags/career-advice.md>), [conversations](<https://devfeed.tech/tags/conversations.md>), [data](<https://devfeed.tech/tags/data.md>), [employees](<https://devfeed.tech/tags/employees.md>), [in-brief](<https://devfeed.tech/tags/in-brief.md>), [messages](<https://devfeed.tech/tags/messages.md>), [meta](<https://devfeed.tech/tags/meta.md>), [mixing](<https://devfeed.tech/tags/mixing.md>), [muse](<https://devfeed.tech/tags/muse.md>), [orders](<https://devfeed.tech/tags/orders.md>), [private](<https://devfeed.tech/tags/private.md>), [product](<https://devfeed.tech/tags/product.md>)

### AI overview

Meta employees reportedly raised concerns that its AI agent Muse mixed up conversations across accounts and exposed private information. Examples in internal posts included unrequested childcare and career messages, product orders, bank balances, and a reminder about a Signal message that one employee said she had not authorized Muse to access. The account is based on a Reuters report of internal posts; Meta had not immediately responded to TechCrunch's request for comment. The excerpt also says Muse launched in September and had more than 10 million downloads.

### Source excerpt

In some of these instances, employees said Muse surfaced messages about childcare requests they never made, career advice they never asked for, product orders they never placed, and bank account balances that didn't belong to them.

## Who controls your site? Securing AI agent access in WordPress

DevFeed: [Who controls your site? Securing AI agent access in WordPress](<https://devfeed.tech/articles/who-controls-your-site-securing-ai-agent-access-in-wordpress-145238.md>)

Original publisher: [Read original article](<https://kinsta.com/blog/control-ai-agent-access-in-wordpress/>)

Author: Carlo Daniele

Published: 2026-11-19T13:30:37Z

Content type: article

Language: en

Sources: [WordPress Guides, Tutorials & Tips - Kinsta® Blog](<https://devfeed.tech/sources/wordpress-guides-tutorials-tips-kinsta-blog.md>)

Topics: [WordPress](<https://devfeed.tech/topics/wordpress.md>), [Access Control](<https://devfeed.tech/topics/access-control.md>), [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [Zero Trust](<https://devfeed.tech/topics/zero-trust.md>), [WordPress AI](<https://devfeed.tech/topics/wordpress-ai.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [ai-agent-access](<https://devfeed.tech/tags/ai-agent-access.md>), [controls](<https://devfeed.tech/tags/controls.md>), [featured](<https://devfeed.tech/tags/featured.md>), [image](<https://devfeed.tech/tags/image.md>), [securing-ai](<https://devfeed.tech/tags/securing-ai.md>), [site](<https://devfeed.tech/tags/site.md>), [wordpress](<https://devfeed.tech/tags/wordpress.md>)

### AI overview

The article's central finding is that WordPress agent safety must rely on limits outside the model, since comments and other external content can carry indirect prompt injections. For comment moderation, expose a narrow tool that accepts a comment ID and valid status, validate it before execution, and use a WordPress identity limited to moderating comments. Keep credentials in a connector, scope access to the correct client site, minimize data sent to the model, and require expiring, environment-specific approval for higher-risk actions. Test rejected values, cross-site IDs, and staging approvals that must not work in production; log rejection reasons without storing passwords or comment text. Kinsta provides site container isolation, managed access, staging, and backups, but these do not limit an API key's authority or prevent data already sent elsewhere from leaking. The agency remains responsible for credentials, WordPress capabilities, tool exposure, and escalation procedures.

### Source excerpt

Featured image for Who controls your site? Securing AI agent access in WordPress

[Next page](<https://devfeed.tech/topics/ai-agent.md?cursor=WyIyMDI2LTExLTE5VDEzOjMwOjM3KzAwOjAwIiwgIjdmOTBiMTRlLTVlNzYtNDQ3OC1hZTk2LTM1NGUzZTE0NDI1MCJd>)