Three-file agent memory design: fixed budgets, details must earn their place
eyishazyer · x · 2026-11-04
Against the common instinct that agents should "remember more," the author lays out a three-layer, file-based memory design:
- recent.md: full recent context, with older content rolling off.
- state.md: one summary of where things stand, rewritten every session, never appended.
- anchors.md: hard cap of 30 facts — decisions, constraints, names.
The core idea (EdgeRock's approach): keep a fixed memory budget and make every detail earn its place, rather than accumulating context indefinitely.
More from coding & agent
- Axion 1.0: ByteDance-Backed Open Agent Foundation Model With 3-Stage Recipe — Yuqi Jia · 2026-11-04
- Caper Pre-Executes Shell Commands in Sandboxes for Safer Agent Approvals — SingaporeManagementUniversity · 2026-11-04
- Claude Code lead: AI shifts team bottlenecks upstream to PMs and downstream to verification — Flomerboy · 2026-11-04
- 10 open-source repos that run AI agents while you sleep: cron, webhooks, cloud PRs — victor_explore · 2026-11-04
- The shopping cart pattern: recursively chaining AI micro-decisions to handle multi-intent queries — hardimanjames · 2026-11-04
- Researcher: Codex is 'miles ahead' of Claude Code on computer use — Yuchenj_UW · 2026-11-04