{"id":38975,"date":"2026-08-21T15:27:40","date_gmt":"2026-08-21T15:27:40","guid":{"rendered":"https:\/\/lamarr-institute.org\/publication\/environment-grounded-automated-prompt-optimization-for-llm-game-agents\/"},"modified":"2026-09-17T11:17:27","modified_gmt":"2026-09-17T11:17:27","slug":"environment-grounded-automated-prompt-optimization-for-llm-game-agents","status":"publish","type":"publication","link":"https:\/\/lamarr-institute.org\/de\/publication\/environment-grounded-automated-prompt-optimization-for-llm-game-agents\/","title":{"rendered":"Environment-Grounded Automated Prompt Optimization for {LLM} Game Agents"},"content":{"rendered":"<p>{LLM} agents in interactive environments are highly sensitive to their prompts, yet prompt engineering remains a manual, task-specific process. We introduce an automated prompt optimization framework for {LLM} agents that decomposes the observation-to-action pipeline into a goal-conditioned descriptor agent and an action selection agent, and iteratively refines each module&#8217;s prompt through an {LLM}-driven evolutionary loop guided by environment returns. We propose a behavior analyzer to attribute episode outcomes to specific prompt components, and a mutator to propose targeted revisions to the prompt, before validating them through environment rollouts. We evaluate on all five {BabyAI} tasks in the {BALROG} benchmark, comparing our pipeline against {BALROG}&#8217;s {RobustCoTAgent} under both plain and guided prompt initializations. Optimization improves performance consistently across tasks and conditions, without requiring updates to the model weights. On {PutNext}, a multi-step coordination task where the {RobustCoTAgent} achieves 0\\% success, our framework reaches up to 72.5\\% success rate using the same underlying {LLM} with optimized prompts. These results suggest that a multi-agent framework, combined with automatic prompt optimization, enhances {LLMs} without the need for fine-tuning or extensive human supervision.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>{LLM} agents in interactive environments are highly sensitive to their prompts, yet prompt engineering remains a manual, task-specific process. We introduce an automated prompt optimization framework for {LLM} agents that decomposes the observation-to-action pipeline into a goal-conditioned descriptor agent and an action selection agent, and iteratively refines each module&#8217;s prompt through an {LLM}-driven evolutionary loop guided by environment returns. We propose a behavior analyzer to attribute episode outcomes to specific [&hellip;]<\/p>\n","protected":false},"author":14,"featured_media":0,"template":"","meta":{"_acf_changed":false,"footnotes":""},"publication-type":[30],"class_list":["post-38975","publication","type-publication","status-publish","hentry","publication-type-article"],"acf":[],"publishpress_future_workflow_manual_trigger":{"enabledWorkflows":[]},"_links":{"self":[{"href":"https:\/\/lamarr-institute.org\/de\/wp-json\/wp\/v2\/publication\/38975","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/lamarr-institute.org\/de\/wp-json\/wp\/v2\/publication"}],"about":[{"href":"https:\/\/lamarr-institute.org\/de\/wp-json\/wp\/v2\/types\/publication"}],"author":[{"embeddable":true,"href":"https:\/\/lamarr-institute.org\/de\/wp-json\/wp\/v2\/users\/14"}],"version-history":[{"count":0,"href":"https:\/\/lamarr-institute.org\/de\/wp-json\/wp\/v2\/publication\/38975\/revisions"}],"wp:attachment":[{"href":"https:\/\/lamarr-institute.org\/de\/wp-json\/wp\/v2\/media?parent=38975"}],"wp:term":[{"taxonomy":"publication-type","embeddable":true,"href":"https:\/\/lamarr-institute.org\/de\/wp-json\/wp\/v2\/publication-type?post=38975"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}