Anthropic, Effective harnesses for long-running agents

Worauf sich diese Seite beruft, wörtlich, abgerufen am 14.10.2026:

• The Claude Agent SDK is a powerful, general-purpose agent harness adept at coding, as well as other tasks that require the model to use tools to gather context, plan, and execute.bestätigt 10.11.2026
• The core challenge of long-running agents is that they must work in discrete sessions, and each new session begins with no memory of what came before.bestätigt 10.11.2026
• However, compaction isn’t sufficient.bestätigt 10.11.2026
• an initializer agent that sets up the environment on the first run, and a coding agent that is tasked with making incremental progress in every session, while leaving clear artifacts for the next sessionbestätigt 10.11.2026
• We refer to these as separate agents in this context only because they have different initial user prompts.bestätigt 10.11.2026
• The system prompt, set of tools, and overall agent harness was otherwise identical.bestätigt 10.11.2026
• it’s still unclear whether a single, general-purpose coding agent performs best across contexts, or if better performance can be achieved through a multi-agent architecturebestätigt 10.11.2026