Context engineering for long-horizon agents: what breaks and why?

TL;DR: Long-horizon agent failures are not one problem but four distinct degradation modes, and compaction only helps one of them, according to Agentic Fabriq. Summarisation can also preserve poisoned facts while stripping provenance, so the real control question is which context elements must remain immutable and which can be compressed.

Editorial analysis by NHI Mgmt Group, based on content published by Agentic Fabriq: “Context Engineering for Long-Horizon Agents”.

Key questions

Q: What should teams do when a long-running AI agent keeps repeating old actions?

A: Treat that as context distraction, not a general failure of the model.

Q: Why can summarisation make an AI agent less reliable?

A: Because summarisation can strip away provenance and preserve a wrong fact in cleaner form.

Q: How should security teams decide what to keep outside an agent’s summarised context?

A: Keep anything that defines authority or reversibility outside the summarised region.

Practitioner guidance

• Separate durable state from disposable context Keep the task statement, approval state, and action ledger outside any summarisation path so the agent cannot rewrite them during compaction.
• Classify the failure mode before applying compaction Decide whether the problem is poisoning, distraction, confusion, or clash, then apply the matching control instead of summarising by default.
• Clear tool outputs as soon as they are harvested Replace large raw tool dumps with the extracted answer and a pointer to the source so the window keeps signal without preserving unnecessary bulk.

Bottom line: Long-horizon agents fail in multiple distinct ways, so a single summarisation strategy cannot be the default response.

Explore further

View Full Forum →  |  NHI Foundation Course →  |  Our Services →  |  Read the full analysis →