Abstract

Long-running AI agents survive process restarts, failover, live migration, provider substitution and connectivity loss. Their external operations do not stop when the process does:

  • a payment may already have been accepted;
  • a callback may arrive at the old endpoint;
  • a second runtime may still be active;
  • a quota reservation may be missing from the snapshot;
  • the replacement tool may have different side effects.

A restored checkpoint, a new lease or a received callback is therefore not evidence that an operation did not happen, that it is final, or that the new runtime may act.

This disclosure describes a vendor-neutral, non-actuating method for reconstructing recovery state with evidence:

  • The original mandate is kept separate from execution state.
  • Task, instance, generation, logical operation, attempt and external transaction identities are tracked separately.
  • Outstanding operations and their exposure (spent nonces, reserved quota, pending liabilities) are listed before handoff.
  • A new, never-reused execution generation is fenced at the actual effect boundary, in the manner of fencing tokens, so a stale owner cannot cause effects.
  • Callbacks to old or new endpoints are authenticated, correlated and deduplicated as historical evidence, without granting ownership.
  • Authorization use and quota are reconciled across regions, so an unknown reservation is never restored as available.
  • Resuming requires current authority, an equivalent tool from the provider and proven retry safety.
  • Under connectivity loss, only bounded work with no external effect continues.

Two separate three-valued verdicts are produced, one for continuation evidence and one for external-outcome evidence. An unknown outcome is never rewritten as success, failure or cancellation.

The disclosure provides:

  • a 48-stage, six-phase workflow with gated flowcharts;
  • six embodiments;
  • a 20-page mathematical specification (J-M01 to J-M52) with conditional proofs and counterexamples;
  • 40 checkpoints;
  • a comparison with durable workflow engines, CRIU and VM live migration, leader leases and fencing tokens, consensus, idempotency keys, webhook signing, distributed rate limiting, API versioning and saga patterns;
  • a non-actuating evaluator in Python, JavaScript and Go with 122 shared synthetic vectors, all passing (about 0.5 to 1.6 microseconds median per in-memory evaluation).

It is not a consensus protocol or an enforcement mechanism. The laboratory is synthetic, self-authored and not peer reviewed.

A companion industry-relevance section aligns the method with public roadmaps:

  • Amazon Web Services: Bedrock AgentCore Runtime sessions of up to 8 hours on microVMs and up to 14 days on runtime instances.
  • Google: Vertex AI Agent Engine sessions and Memory Bank.
  • Other durable-agent platforms: Microsoft Foundry hosted agents and Azure Durable Functions, Temporal (used with OpenAI Codex and the OpenAI Agents SDK), LangGraph, DBOS and Inngest.
  • Chinese agent-runtime platforms: Alibaba Cloud Agent Bay, ByteDance Volcano Engine ArkClaw, Tencent Cloud, Huawei Cloud, Baidu AI Cloud and Zhipu AI.

Disclaimer: The industry and product-roadmap relevance is an AI-generated alignment based only on public announcements, documentation and press coverage. It is not an endorsement, partnership, sponsorship, licence, technical validation or evaluation by any company or standards body named. It does not state or imply that any company uses, needs, has assessed or infringes this method or any patent application. All company and product names are trademarks of their respective owners and are used only to describe technical relevance. Patent pending: Indian Patent Application No. 202631129178 (source disclosure on Zenodo, record 23793543). Related PCT application: PCT/IB2026/055615, filed 4 June 2026 and publicly disclosed 16 July 2026. Readers must read the original disclosures before drawing any conclusion. The inventor is open to strategic collaboration or acquihire discussions.

Keywords

AI safety, persistent agents, long-running agents, durable execution, runtime migration, failover, checkpoint and restore, fencing tokens, execution generation, split-brain, asynchronous callbacks, webhook correlation, idempotency, retry safety, distributed quota, authorization continuity, provider compatibility, partition tolerance, three-valued logic, Amazon Bedrock AgentCore, Vertex AI Agent Engine, Microsoft Foundry, Azure Durable Functions, Temporal, OpenAI Agents SDK, Alibaba Agent Bay, Volcano Engine

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial 4.0 License

Download

Share

COinS