Mining a directly-follows workflow graph from black-box conversations with an LLM agent enables structurally targeted boundary testing that covers 23–38 distinct stateful boundaries per agent, roughly doubling prompt-only baselines.
Title resolution pending
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.SE 2years
2026 2representative citing papers
An evidence-based promotion/demotion lifecycle converts validated LLM agent traces into zero-token deterministic workflows, reducing per-incident cost by 70% in a production cloud-networking system.
citing papers explorer
-
Mining Workflow Graphs for Black-Box Boundary Testing of Conversational LLM Agents
Mining a directly-follows workflow graph from black-box conversations with an LLM agent enables structurally targeted boundary testing that covers 23–38 distinct stateful boundaries per agent, roughly doubling prompt-only baselines.
-
Progressive Crystallization: Turning Agent Exploration into Deterministic, Lower-Cost Workflows in Production
An evidence-based promotion/demotion lifecycle converts validated LLM agent traces into zero-token deterministic workflows, reducing per-incident cost by 70% in a production cloud-networking system.