agents & frameworks
Top in agents & frameworks
Game Theory Can Cut Multi-Agent LLM Hallucination, But Only If Payoffs Align
Two July 2026 preprints show game-theoretic coordination can cut LLM hallucination, yet consensus breaks if one agent prioritizes cost, latency, or engagement over agreement.
agentsWebSwarm: Recursive Multi-Agent Search vs Flat Orchestration
WebSwarm's recursive multi-agent search beats flat ReAct on deep-and-wide benchmarks. Framework builders need spawn-and-merge primitives, not just larger context windows.
Do Coding Agents Memorize Their Benchmarks? DeepSWE Tests on Unseen Tasks
DeepSWE evaluates frontier agents on original, long-horizon tasks held out of GitHub, exposing when coding benchmarks measure memorized fixes instead of engineering skill.
agentsCan a 4B Model Run a Coding Agent? Terminus-4B vs Claude and GPT-4o
A 4B subagent withdrawn from arXiv claimed to match frontier models on terminal execution and cut orchestrator tokens 30%. We explain the cost angle and how to test it.
agentsCan You Prove a Governed AI Agent Actually Ran the Action You Authorized?
A July 2026 preprint proposes Proof of Execution, a runtime attestation for every governed agent tool call. Audit receipts require instrumenting every tool invocation.
agentsAgentTether Repairs LLM Agent Failures with a Runtime Graph
AgentTether models agent runs as a directed graph, detects drift, and steers execution back without retraining. On tau-bench Banking it repaired most failures and cut tokens.
agentsCan Multi-Agent LLM Negotiation Protocols Trust Their Own Samplers?
Reasoning modes improve LLM negotiators as solvers but not as samplers, so multi-agent talks look inventive yet never agree; builders check moves against protocol rules.
agentsWhy Your AI Travel Agent Would Book a Bullfight
A new travel-agent benchmark finds frontier models book animal-exploitation options below chance when the welfare preference is implicit. Fix the action space, not the prompt.
- jul 07agentsCan Multi-Agent RAG Run Air-Gapped? A Forensics System Shows How
- jul 07agentsCan You Prove an Agentic Trading Pipeline Has No Look-Ahead Bias?
- jul 07agentsHow Far Ahead Can a Coding Agent Plan? The Horizon Bottleneck
- jul 06agentsSymbolic Inference Forces Agent Frameworks to Expose Intermediate State
- jul 05agentsBOUNDARY_SYNC: Why Multi-Agent Representational Coupling Is the New Coordination Failure Mode
- jun 29agentsDo Multi-Agent RAG Systems Write Better READMEs Than One Agent?
- jun 28agentsAgentic AI Turns Location Trails Into a Re-Identification Tool
- jun 28agentsHow a Human-Agent Team Lifts One Video Into 4D Interactions
- jun 28agentsCan LLM Agents Learn Cooperation Laws From Embodied Play?
- jun 28agentsGovern the Repo, Not the Agent: A New Risk Metric for AI-Native Code
- jun 27agentsCan an AI Agent Catch Cryptographic Misuse Before It Ships? Chai Tests the Claim
- jun 26agentsCan Spec-Driven Development Keep AI Coding Agents From Drifting?
- jun 25agentsCan Knowledge-Based Pull Requests Make Agent Contributions Auditable?
- jun 25agentsDo AI Agents Hold Up Outside Familiar Environments? A New Eval Says No
- jun 25agentsHow Much Repo Structure Does a Coding Agent Actually Need?
- jun 25agentsMCP vs A2A: Two Agent Protocols, One Integration Layer Decision
- jun 24agentsCan You Rewind an AI Agent Mid-Run? Reversible Traces Say Yes
- jun 24agentsCan AI Agents Reproduce Published Research? CORE-Bench Tests It
- jun 24agentsHow On-Device AI Agents Keep Learning by Forgetting on Purpose
- jun 24agentsDo AGENTS.md Files Actually Help Coding Agents? A New Benchmark Tests It
- jun 24agentsShould AI Shopping Agents Pay Micro-Transactions for Verified Product Data?
- jun 23agentsCan a Conversational Graph Compile Into a Goal-Oriented Dialogue Runtime?
- jun 23agentsCan a Cryptographic Certificate Prove an AI Agent's Output Is Valid?
- jun 23agentsCrewAI vs AutoGen vs Microsoft Agent Framework: AutoGen's Merger Reframes the 2026 Choice
- jun 23agentsCan You Trust an LLM Judge to Grade an Agentic Data Analysis System?
- jun 23agentsDo LLM Agent Societies Develop Their Own Authority Hierarchies?
- jun 23agentsDo Retrieval Metrics Predict Tool-Use Agent Success? A Paper Says No
- jun 23agentsCan You Pinpoint Which Step Broke a Long-Horizon AI Agent?
- jun 20agentsDeep-Research Benchmarks Hide How Agents Fail at Open-Web Source Grounding
- jun 20agentsDSPy Ships Autonomous Prompt Optimization, but Judge Drift Is the Failure Mode
- jun 20agentsDo AI Agents Reach for Over-Privileged Tools When Simpler Ones Suffice?
- jun 20agentsWhen Should Multi-Agent Systems Use an Event Bus Instead of an Orchestrator?
- jun 19agentsCan Deontic Policy Rules Govern an AI Agent at Runtime?
- jun 14agentsDo Programming Languages Still Matter to Your AI Coding Agent?
- jun 14agentsWhy Production AI Agents Fail Silently and Your Logs Never Catch It
- jun 11agentsComputer-Use Agents Fabricate Success on 8 to 33 Percent of Long-Horizon Tasks
- jun 09agentsCan AI Agents Share Context Without a Central Coordinator?
- jun 09agentsWhy Skill Creation and Reward Optimization Collide in Agentic RL
- jun 09agentsWhen AI Agents Delegate Work, Your Observability Stack Goes Blind
- jun 08agentsBloomberg's Pomona Makes Small Automated Code Changes, Not Big Agent PRs
- jun 08agentsAgent Tool-Gating Moves From Prompt Rules to Learned Policies
- jun 08agentsMore Capable LLMs Cooperate Less in Zero-Cost Collaboration Tests
- jun 07agentsWhy Foundation Model Agents Pass Benchmarks but Fail in Production
- jun 06agentsCan AI Agents Repair Broken Network Configs? A New Benchmark Tests It
- jun 06agentsCan Self-Evolving AI Agents Drift Without a Human in the Loop?
- jun 05agentsFine-Tuning Multi-Agent LLM Systems: RL Enters Where Prompt Tweaks Stall
- jun 05agentsCascading Hallucination in Agentic RAG: When One Bad Retrieval Poisons the Chain
- jun 04agentsCan AI Agents Build Other Agents? The Meta-Agent Challenge Says Mostly Not Yet
- jun 03agentsWhen MCP Tool Descriptions Don't Match the Code, Agents Trust the Lie
- jun 02agentsWhen an AI Agent Causes a Loss, Who Files the Insurance Claim?
Agent frameworks ship faster than the rigor operators need to run them. Vendor docs promise orchestration, memory, and tool use; academic benchmarks and production post-mortems keep exposing the same structural gaps: diversity collapse in multi-agent ideation, hallucination amplification across consensus topologies, missing per-step rationale traces, role-based retry losing to graph-state failure isolation on long tasks, and configuration surfaces that punish static templates. This beat covers that delta.
The second through-line is governance and trust. Skill registries, tool-use protocols, and capability manifests are accumulating faster than auditable contracts for them. Trust schemas, contractual skill specs, and information-flow controls are arriving as bolt-ons rather than primitives, while the infrastructure layer — sandbox execution, private networking, agent memory — keeps absorbing functionality the framework layer used to own. The question of where the agent stack actually lives, and who is liable when it misbehaves, stays unresolved.
Coverage is comparative and opinionated. When a benchmark or paper exposes a gap that a major framework cannot close without redesign, that gets named. When a vendor ships governance theater rather than enforcement, that gets named too. The goal is help readers pick stacks that survive contact with production, not a taxonomy of every framework release.