Context Hygiene for Coding Agents: Murder, Lobotomy, or Diary
A million-token context window doesn’t mean you should fill it.
Every piece of information you keep in a coding agent’s active context has a cost. Not just in tokens and dollars, but in the model’s attention.
In this episode, we explore context hygiene for coding agents: how to decide what deserves to stay in active context, what to move elsewhere, and when to throw context away entirely.
00:00 Intro: context hygiene
02:26 Why context size matters
03:22 Discussing Steve Yegge: sentience & encouragement
11:27 How context and prompt caching work
18:23 Compacting vs. clearing context
19:56 The lazy approach: automatic compaction
24:28 The high-touch approach: handoffs before clearing
30:34 Automating context handoffs with Pi
34:51 /btw, /side, and Pi’s more powerful tree branching
39:55 Offloading work to sub-agents
41:31 Moving coordination into Beads and durable state
Superlinear is a podcast about emerging practices for building with coding agents.
Hosted by Brandon Kase and Christine Yip.
Homepage: https://superlinear.fm
Twitter: https://x.com/superlinear_fm
LinkedIn: / superlinearfm
Brandon Kase: https://x.com/bkase_
Christine Yip: https://x.com/christinetyip
Music licensed through Soundstripe.
Code: O2UT6CCEQNH4KT0D