index 56d4a2f..4a6661d 100644
@@ -1,7 +1,13 @@
---
+first_captured: 2026-04-10
+sources:
+- sources/google-sheets-ideas.md
status: raw
tags:
- ai
+- agents
+- context-engineering
+- software
title: context window optimizer
type: idea
updated: 2026-04-11
@@ -10,4 +16,8 @@ visibility: public
# context window optimizer
-tooling for agent context management.
\ No newline at end of file
+tooling for agent context management — the problem of deciding what goes into a limited context window when agents work with large, complex knowledge bases. most agents either stuff everything in (hitting the limit, degrading quality) or use naive retrieval (top-k semantic search, which misses structural dependencies). neither approach is good. the optimizer sits between the knowledge base and the agent, managing what gets surfaced, in what order, at what point in the task.
+
+the technical research here is genuinely interesting: context window management involves tradeoffs between recency, relevance, diversity, and dependency structure. a code editing agent needs the current file, related interfaces, recent change history, and relevant documentation — and those come from different retrieval strategies. the 40% context utilization target in [[spec-driven-dev|spec-driven dev kit]] is a practical heuristic: leave room for the agent's own output and reasoning, otherwise you get truncated context that causes errors mid-task. good tooling would make this principled: profile the task type, estimate context needs, and dynamically manage the budget across a multi-step workflow.
+
+this is infrastructure for the [[axon|axon]] vision — once you have a rich personal context store, you need a smart layer that decides what to surface to any given agent call. it also connects directly to [[cognitive-foom|cognitive foom]] for the self-improvement angle: an agent that gets better at its own context management is effectively improving its own reasoning capacity. the [[cluster-ai-tools|AI tooling]] cluster treats this as connective tissue — it is not a product on its own but enables every other agent-based product to work better. [[hard-docs-writer|hard docs writer]] is a related idea: documentation written specifically to be machine-readable is the complement to a context optimizer that knows how to use that documentation.
\ No newline at end of file