work-operating-model

An operating model interview maps your tacit work patterns across five layers: operating rhythms, recurring decisions, domain encoding, dependencies, and artifacts. It extracts institutional knowledge and friction points through conversational elicitation using recent concrete examples.

Start a work operating model interview session by using the start_operating_model_session MCP recipe tool. You provide recent concrete examples from your last week, and the skill sequentially guides you through five ordered layers of work pattern elicitation.

Exporting an operating model generates operating-model.json, USER.md, SOUL.md, HEARTBEAT.md, schedule recommendations, and ARTIFACTS.md. These files provide structured agent-ready outputs for cross-tool knowledge reuse.

Yes, the operating model interview requires base Open Brain search and capture tools. It also relies on MCP recipe tools like start_operating_model_session and generate_operating_model_exports to save layers and produce final outputs.

Yes, you can check for contradictions across operating model layers before exporting. The skill runs consistency checks across all five approved layers and requires explicit user confirmation through checkpointed summaries before saving any data.

The best way to document institutional knowledge for AI agents is through a structured five-layer interview that captures behavioral relationships and domain encoding. This converts tacit experience into persistent files like SOUL.md and operating-model.json for agent reuse.