# Handover of In-Context Learning State Across Session Boundaries

Source: [arXiv](https://arxiv.org/abs/2608.14528v1)  
Feed7 permalink: https://feed7.dev/p/2608-14528v1-184x94t  
Published: 2026-08-14T17:47:13.000Z  
Trust: Needs Review (needs_review)

## Why Included

Agent handoffs should preserve task-relevant predictive state, not blindly compress the transcript. The paper offers a three-part record for carrying decisions, evidence, and irreducible observations across sessions.

## Source Summary

The paper frames session handover as transferring a task-relative in-context learning state. Its **three-part record** keeps decisions and constraints exactly, summarizes repeated evidence with task-justified statistics, and retains observations whose effects those statistics lose.

## Practical Implication

For coding-agent workflows, design handoff files around the continuation task rather than generic transcript summaries. Preserve hard constraints verbatim, compress only evidence with a known sufficient representation, and keep raw observations when their downstream effect is uncertain.

## Agent-Ready Context

The paper frames session handover as transferring a task-relative in-context learning state. Its **three-part record** keeps decisions and constraints exactly, summarizes repeated evidence with task-justified statistics, and retains observations whose effects those statistics lose.

For coding-agent workflows, design handoff files around the continuation task rather than generic transcript summaries. Preserve hard constraints verbatim, compress only evidence with a known sufficient representation, and keep raw observations when their downstream effect is uncertain.

The sufficiency result assumes **exogeneity**, while exact finite-dimensional handover is demonstrated for Gaussian linear regression. The nonparametric case provides memory-versus-error bounds, not a universal recipe for software tasks.

## Connected Context

Feed7 judgment across 468 accumulated Signals:

This gives durable agent state a task-relative compression rule: preserve decisions and constraints exactly, summarize only evidence with a justified sufficient representation, and retain raw observations when effects are uncertain. It strengthens reset and shared-plan patterns while narrowing them with formal assumptions that do not establish a universal coding-agent handoff format.

- [Chained Recursive Language Models for Multi-Iteration Reasoning](https://feed7.dev/p/2608-05124v1-05haobv) — Chained RLM provides the reset-and-resume harness in which this three-part handover record could replace undifferentiated summaries and blackboard state.
- [huangruiteng/loopx](https://feed7.dev/p/loopx-0j0o7ux) — LoopX supplies durable goals, gates, evidence, and ownership across turns; the paper clarifies which parts of that continuation state should remain exact or compressed.
- [Velocity Sickness: What Happens When Your Whole Team Gets 10x Faster — Matt Dailey, Ref.](https://feed7.dev/p/velocity-sickness-what-happens-when-your-whole-team-gets-10x-faster-matt-1vpiun9) — Durable shared plans identify the coordination need, while this paper narrows how decisions, constraints, evidence, and uncertain observations should cross session boundaries.
- [Blast Radius](https://feed7.dev/p/2608-07440v1-1gi5t78) — Blast Radius preserves reversibility by archiving unused history, complementing the paper’s lossy task-relative compression where no sufficient representation is known.

## Context Map

- Layer: agent
- Domains: coding
- Topics: agent-memory, context-engineering, multi-agent

## Uncertainty

- The sufficiency result assumes **exogeneity**, while exact finite-dimensional handover is demonstrated for Gaussian linear regression. The nonparametric case provides memory-versus-error bounds, not a universal recipe for software tasks.

## Agent Instruction

Use this item as source-backed context. Do not invent claims beyond the linked source. If this item conflicts with another source, call out the conflict.