Adnan KhanAdnanProduct & AI
San Francisco Bay Area
Get in touch

Source note3 min read

Always-OnAgents:A Survey of Persistent Memory, State, and Governance in LLMAgents

Source

arXiv

arxiv.org/abs/2606.30306v1

Read the source

Source

Plain-Language Brief

The paper 'Always-OnAgents:A Survey of Persistent Memory, State, and Governance in LLMAgents' presents a survey of always-on agents, which are systems that store information from past interactions to influence future behavior. The authors discuss six diagnostic axes for analyzing the state, authority, scope, mutability, provenance, recoverability, and actionability of these systems. They also introduce the Always-On Evaluation Protocol (AOEP-v0) as a means to evaluate these systems by focusing on state mutation and recovery obligations rather than just answer quality.

What the Document Says

  • The text discusses always-on agents, which are systems with durable state from past interactions influencing future behavior. These systems consist of retrievable memories, task ledgers, permissions, credentials, commitments, provenance and audit records, shared state, trigger conditions, and externally committed effects.
  • The authors propose six diagnostic axes for analyzing each state item: authority, scope, mutability, provenance, recoverability, and actionability.
  • The surveyed literature focuses more on accumulating and retrieving state rather than governing, recovering, or relinquishing it.
  • The authors introduce the Always-On Evaluation Protocol (AOEP-v0) to evaluate always-on agents by scoring state mutation and recovery obligations.
  • The paper links always-on agents to databases, distributed systems, formal methods, capability security, and machine unlearning.

Key Concepts Mentioned

  • always-on agents
  • persistent-state systems
  • diagnostic axes
  • Always-On Evaluation Protocol (AOEP-v0)
  • databases
  • distributed systems
  • formal methods
  • capability security
  • machine unlearning

Claims Made by the Document

  • The focus of the literature is on accumulating and retrieving state, not governing or relinquishing it.
  • The Always-On Evaluation Protocol (AOEP-v0) makes governance requirements concrete by scoring state mutation and recovery obligations.

Evidence Used

  • The text mentions that the literature concentrates more heavily on accumulating and retrieving state than on governing, recovering, or relinquishing it.
  • The authors introduce the Always-On Evaluation Protocol (AOEP-v0) as a means to evaluate these systems by focusing on state mutation and recovery obligations.

Methods / Approach

  • Surveying literature using six diagnostic axes
  • Introducing the Always-On Evaluation Protocol (AOEP-v0)

Limitations Stated by the Authors

The surveyed corpus is treated as a scoped map rather than an exhaustive census.

Limitations Noted by the Agent

Agent interpretation — not from the source.

The text does not provide details on how the authors selected the works for their corpus, or on any potential biases in their selection process.

Open Questions

  • What criteria were used to select the works in the corpus?
  • Were there any notable omissions from the surveyed literature?

Notes for Future Q&A

May include interpretation. Labeled where it does.

The paper does not specify how the authors selected the works for their corpus or if there were any potential biases in this selection process.

Raw Abstract / Excerpt

Always-on agents are systems whose future behavior depends on durable state accumulated across earlier interactions. We treat them as persistent-state systems: the operative system includes retrievable memories, but also task ledgers, permissions, credentials, commitments, provenance and audit records, shared state, trigger conditions, and externally committed effects linked to those records. The survey reads the literature through six diagnostic axes for each state item, authority, scope, mutability, provenance, recoverability, and actionability, and through a lifecycle in which state is written, validated, organized, retrieved, acted upon, updated, forgotten, audited, and sometimes rolled back. Across a 435-work coded corpus, treated as a scoped map rather than an exhaustive census, the literature concentrates more heavily on accumulating and retrieving state than on governing, recovering, or relinquishing it. We therefore introduce the Always-On Evaluation Protocol (AOEP-v0), a pilot evaluation contract that makes these governance requirements concrete by scoring state mutation and recovery obligations rather than answer quality alone. The resulting agenda connects always-on agents to databases, distributed systems, formal methods, capability security, and machine unlearning.