Which of the following are valid strategies for mitigating context pollution in a long-running agent?

Attempt it

Quick reference

TL;DR

Real mitigations actively shrink or filter the context: sliding window, summarisation, and pruning. A bigger window or disabling retries does not fix the underlying noise.

Memory aid

Sign in for the mnemonic that makes this stick. See Pro pricing.

Easy to grasp

Imagine a detective's desk that fills with notes as a case drags on. Every lead, every dead end, every coffee-stained scribble piles up. Soon the desk is so cluttered the detective wastes time re-reading junk and misses the one note that matters. To stay sharp they can keep only the newest pages, rewrite the old ones into a tidy summary, or throw out notes that turned out to be irrelevant. What does not help is buying a bigger desk. The clutter just spreads to fill it, and the useful fraction stays just as low. An agent's context window is that desk. The fix is to manage what sits on it, not to make it larger.

Key concepts

Context Management Context Pollution Summarisation

Quick scan · 2–4 min

Everything important, quickly.

Define pollution as a noise ratio problem, name the three active mitigations with one tradeoff each, explain why a bigger window and disabling retries fail the test, then close with how production tiers these techniques and adds sub-agent isolation.

Side-by-side

Strategy	Mechanism	Main cost	Genuine mitigation
Sliding window	Keep recent N steps	Forgets early constraints	Yes
Summarisation	Compress old steps every K turns	Extra LLM call, lossy	Yes
Selective pruning	Drop low-relevance entries	Needs a relevance signal	Yes
Bigger window	More capacity	Noise still accumulates	No
Disable retries	Fewer failed calls stored	Breaks error recovery	No

Real-world examples

Real products, models, and research that use this idea.

Claude Code compacts long sessions by summarising earlier turns into a condensed memory block once the transcript nears the window limit, then continues with that summary plus recent turns.
LangGraph exposes a state reducer and trimming hooks so a ReAct loop can apply a sliding window or summarise old messages between iterations.
Anthropic's Claude Opus 4.7 supports very large context, yet its own agent guidance still recommends summarisation and offloading because raw recall decays in the middle of long inputs.
Cursor and Cline store full file contents and tool outputs outside the prompt, injecting only the slices the current edit needs to keep the coding agent's context lean.

Interviewer follow-ups

What an interviewer would ask next. Try answering before peeking at the approach.

QHow would you protect the original goal and hard constraints when using a sliding window that forgets early turns?

Sign in or go Pro to see the approach for this follow-up.

QWhen does summarisation hurt more than help in an agent loop?

Sign in or go Pro to see the approach for this follow-up.

QHow does sub-agent isolation address pollution differently from windowing or pruning?

Sign in or go Pro to see the approach for this follow-up.

60-second cheat sheet

The night-before-the-interview bullets. Scan these on the way to the call.

Sign in for 5 unlocks/day or go Pro for unlimited revision sheets. See Pro pricing.

References

Primary sources. Skim if you want the original framing.