AWS's Strands Evals SDK and Amazon Bedrock AgentCore Evaluations add skill-focused evaluators for agents that use modular SKILL.md skills. They separate two failure modes general metrics miss: wrong skill selection and incomplete instruction following, using recorded trajectories or OpenTelemetry traces.
AWS published a technical walkthrough of evaluating multi-agent systems using Amazon Bedrock AgentCore Evaluations with built-in and custom evaluators. The reference implementation is a supply chain decisioning system with an orchestrator and four specialized sub-agents built on Strands Agents SDK, showing how teams can measure accuracy, business validity, and explainability from development through production.
AWS published a technical walkthrough of evaluating multi-agent systems using Amazon Bedrock AgentCore Evaluations with built-in and custom evaluators. The reference implementation is a supply chain decisioning system with an orchestrator and four specialized sub-agents built on Strands Agents SDK, showing how teams can measure accuracy, business validity, and explainability from development through production.
Anthropic introduced Agent Skills, an open specification for extending LLM agents with reusable procedural knowledge loaded on demand. A new arXiv note argues skills should be authored as software artifacts, applying software-engineering principles like single responsibility and behavioral evaluation.
Anthropic introduced Agent Skills, an open specification for extending LLM agents with reusable procedural knowledge loaded on demand. A new arXiv note argues skills should be authored as software artifacts, applying software-engineering principles like single responsibility and behavioral evaluation.
Researchers introduced Skill-Use, a benchmark evaluating whether LLM agents can recognize relevant skills and apply them under progressive disclosure, where only a skill's name and short description are shown initially. It separates three facets of skill use, addressing a gap in existing evaluations that focus on skill quality or task success rather than autonomous application.
Researchers introduced Skill-Use, a benchmark evaluating whether LLM agents can recognize relevant skills and apply them under progressive disclosure, where only a skill's name and short description are shown initially. It separates three facets of skill use, addressing a gap in existing evaluations that focus on skill quality or task success rather than autonomous application.
An agentic platform AWS is building with Wood Mackenzie on Amazon Bedrock AgentCore, addressing the gap between agent prototypes and production deployments. It standardizes concurrency, session isolation, identity, state, scaling, and guardrails so the energy-research firm doesn't rebuild this layer per agent. LLM-based agents run on real traffic with monitoring to verify correct behavior.
An agentic platform AWS is building with Wood Mackenzie on Amazon Bedrock AgentCore, addressing the gap between agent prototypes and production deployments. It standardizes concurrency, session isolation, identity, state, scaling, and guardrails so the energy-research firm doesn't rebuild this layer per agent. LLM-based agents run on real traffic with monitoring to verify correct behavior.
Yogendra Miraje of FactSet described a skill-centric agent harness where skills replace traditional features. Each skill is defined by a short skill.md file with distinct name and description for routing. The system uses progressive disclosure and trigger words to load only needed skills.
Yogendra Miraje of FactSet described a skill-centric agent harness where skills replace traditional features. Each skill is defined by a short skill.md file with distinct name and description for routing. The system uses progressive disclosure and trigger words to load only needed skills.