Synthesizing Procedural Memory: Challenges and Architectures in Automated Workflow Generation
This work addresses the challenge of enabling AI agents to autonomously generate executable workflows, which is incremental as it builds on the CodeMem paradigm but introduces novel architectural solutions.
This paper tackles the problem of autonomously synthesizing procedural memory for AI agents from scratch, transitioning LLMs from passive tool-users to active workflow architects. Through a case study of a cross-service orchestration task, the authors demonstrate that agents can autonomously write robust, production-grade code skills by addressing four structural bottlenecks in automated skill generation.
While CodeMem establishes executable code as the optimal representation for agentic procedural memory, the mechanism for autonomously synthesizing this memory from a blank slate remains underexplored. This paper operationalizes the transition of Large Language Models from passive tool-users to active workflow architects. Through a high-fidelity case study of a cross-service orchestration task involving Outlook and OneDrive, we identify and address four structural bottlenecks in automated skill generation: the Discovery Gap involving navigation of large tool registries, the Verification Gap regarding grounding tool response structures, the Decomposition Gap which replaces inefficient search with Linear State Anchoring, and the Scaling Gap focused on concurrency and persistence. We demonstrate that by enforcing a scientific methodology of hypothesize, probe, and code, agents can autonomously write robust, production-grade code skills.