tag: llm-context-management
Controlling what information is made available to large language models at invocation time — rules files, system prompts, repository knowledge, and the architecture of what enters (and stays out of) the context window.
- 2026-03-30 — Sub-Agents aren't Agents - They're Skills but Worse
- 2026-03-21 — The Load-Bearing Pipeline Was Human
- 2026-03-14 — The Adeptus Mechanicus Bootcamp: A Gentle Seduction
- 2026-03-09 — Context to Ashes, Skills to Dust
- 2026-02-23 — It's Model Context Protocol, Not Agent Context Protocol
- 2026-02-22 — .gitignore is not .agentignore
- 2026-02-12 — Stop Doing AGENTS.md