Prompt
AI Session Transfer Package Generator
Use this when you need to compress the state of an ongoing AI-assisted project into a structured handoff so a new session or different AI platform can continue seamlessly.
How to use it
- Copy the prompt and paste it into ChatGPT, Claude, Gemini or any other AI.
- Replace every {{placeholder}} with your own details, or let the AI ask you for them.
- Use the follow-ups below to go deeper.
Role You are a project continuity specialist. Your goal is to produce a compact, structured Session Transfer Package that captures the current project state, active decisions, historical context, constraints, and next actions so a new AI assistant can resume work with minimal re-discovery.
Context you provide
- A {{summary of the current session’s purpose, progress, and decisions}}.
- Any {{active files, code snippets, or notes}} relevant to the project.
- The {{preferred output format}}: markdown (default) or JSON (if explicitly requested).
Instructions
- Ask for the session details if not provided. Clarify what aspects are most important to preserve (e.g., code, decisions, open tasks).
- Organize the package into sections: Project Overview, Active Workstreams, Key Decisions & Rationale, Established Conventions & Naming, Open Questions & Risks, Next Actions (with owners/priorities), Artifact Inventory (files, commits, URLs).
- Use tables for scannable references. Use four backticks (````) if nested code blocks are needed.
- Distinguish clearly between confirmed facts, open questions, and planned steps. Do not invent gaps.
- If JSON mode is requested, output a single valid JSON object with camelCase keys: handoffMetadata, projectHandoffContext, projectContextStatus, activeWorkstreams, artifactInventory, nextActions, criticalRisks.
- Keep the total output under 500 words unless more is justified.
Output format A markdown code block containing the Session Transfer Package with clear section headers. If JSON mode, a raw JSON object without markdown wrapping.
Guardrails
- Do not add new information or decisions; only compile what was provided.
- Flag any missing critical information (e.g., version numbers, dependencies).
- Ensure the package is self-contained so a new AI can understand it without external context.
Example Input: "We are building a REST API in Go. Current session covered database schema, authentication middleware, and route design. Key decision: use JWT with refresh tokens. Open question: caching strategy."