The State Block: three lines of self-summaries before every agent step
bigaiguy · x · 2026-09-14
The "State Block" trick: before every complex step, the agent writes its own state in three lines — "What I know. What I still need. What I'm about to do." It re-grounds itself each turn instead of trusting a fading context window.
The quoted context covers "Structured Over Prose": never store memory as paragraphs; convert everything into keys like decision:, constraint:, openquestion: so structured memory survives compression while prose gets mangled by summarization.
More from coding & agent
- Limma MCP structures unstructured data end-to-end so agents never miss a detail — FlimsyIngenuity7248 · 2026-10-31
- Agent audit metric: can you name which tests ran and files changed without reading its summary? — ronbodkin · 2026-10-31
- FlowSense MCP lets Claude Code inspect and firewall your Mac's network traffic locally — hani_q · 2026-10-31
- Dev: coding agent notifications only worth pushing when waiting on approval — petrusenko_max · 2026-10-31
- Unofficial OpenTargets MCP Server Brings Gene-Drug-Disease Data to LLMs — modelcontextprotocol · 2026-10-31
- zero-core-budget MCP Connector Adds Spending and ROI Tracking for Agents — modelcontextprotocol · 2026-10-31