Pollard runtime captures and replays AI agent actions with strict control
Pollard: Governed Semantic Execution Trees for Framework-Independent AI Agent Control and Replay
Artificial IntelligenceMultiagent Systems
Summary
AI agents often perform complex tasks where it's important to track what was done, control resources, and repeat actions exactly. The authors present Pollard, which records each step of an AI agent's decisions and actions in a detailed tree with unique identities, enabling strict replay without repeating external calls. This approach also manages resources carefully across multiple concurrent processes. The authors tested Pollard thoroughly, showing it can verify behavior and avoid errors under specific assumptions.
What this means in practice
- •For ai software engineers: Control and verify AI agent execution with a unified system for recording, replaying, and resource management across frameworks.
- •For cloud infrastructure teams: Ensure exact resource usage and transaction coordination for concurrent AI workloads using Pollard's transactional stores and leases.
Authors
Muntaser Syed, Marius C. Silaghi
Abstract
Agent frameworks can persist checkpoints, trace tool calls, and limit resource use, but these facilities often assign different identities to execution, governance, and replay. We present Pollard, a framework-independent runtime based on governed semantic execution trees. A step identity commits to canonical input, kind, attempt, and parent identity; the same tree records completion, refusal, resource settlement, and branch history. Strict replay returns a verified stored result without invoking the wrapped operation, while a transactional store coordinates reservations across concurrent workers. Under a signed deterministic freeze, two state-model replicas each executed the evaluation transitions without an invariant violation, retained-lease PostgreSQL groups were tested against exact limits, strict replay avoided external calls, and all semantic mutations were detected with no false alarms on the controls. A multi-pair framework matrix preserved reference behavior and verified the trees. A separate cost-capped hosted study exercised live provider paths, and every completed Pollard trial produced a verified tree. Together, the studies support the core mechanisms under explicit lease, boundary, and seal-custody assumptions.