continuity record architect

The problem

Switching between LLMs or starting new chat sessions usually means losing weeks of context. You waste time re-explaining your persona, project status, and the rationale behind previous technical decisions every time you hit a context limit or change providers.

What it does

• Generates a portable Continuity Record in Markdown or YAML to store project identity and stable preferences.
• Maintains a Decision Ledger that tracks not just what was decided, but the specific rationale, confidence levels, and "revisit if" triggers.
• Produces condensed Transition Snapshots and handshake prompts for immediate pasting into new model sessions.
• Resolves information conflicts using a strict hierarchy, prioritizing explicit instructions over agent inferences.
• Tracks "Open Loops" to ensure unfinished tasks and pending dependencies are never forgotten during a handoff.

Why this beats prompting it yourself

Generic prompts for "remembering context" lack a rigorous schema for provenance and decision history. This tool enforces a disciplined structure that separates long-term identity from short-term state, ensuring the next model doesn't just get a wall of text, but a prioritized, actionable manifest of where you left off.

Use cases

• Migrating a complex coding project from one LLM provider to another without losing architectural context.
• Resuming a long-term research task after a multi-week break or session timeout.
• Handing off a specific AI-driven workflow from a developer to a non-technical stakeholder.

Known limitations

This tool produces a document for manual transport. It cannot automatically detect model changes or persist data internally without the user saving the generated file.