Effective harnesses for long-running agents
by Anthropic
Harness design for agents that work across many context windows.
This article addresses continuity when an agent's work spans multiple context windows. It focuses attention on the surrounding execution setup rather than assuming a larger prompt will solve persistence problems.
- Level
- Advanced
- Models
- claude
- Category
- Context Engineering
- Tags
- build-log, harness
- Source date
- 2025-11-26
- Added
- 2026-10-18
- Last verified
- 2026-10-18
- Status
- ok
- Prerequisites
- Requires implementation experience or critical reading of technical methods and system trade-offs.
- Subcategory
- loop engineering
- Region
- West
- Original language
- en
- Access
- free
- Resource reuse
- closed: No general open-reuse grant established from the inspected page. Closed is the conservative index bucket, not a legal finding; the resource is publicly readable. Facet definitions
Summary in our own words. Rights remain with the original creator. Credit and correction policy.