{"slug": "gpt-handover-prompt-md", "title": "GPT_Handover_Prompt.md", "summary": "Christiane Wiemann published a two-step prompt workflow on frauwiemann.de for converting a ChatGPT conversation into a structured Markdown handover document. The first prompt casts the model as a project archivist that extracts 14 sections — including project purpose, key decisions, user preferences, guardrails, open questions and source integrity — while the second prompt runs a quality check on the generated file. The output is intended to be uploaded into a future project or AI environment as compact long-term context.", "body_md": "|  | # GPT Handover Prompt – Two-Step Version | \n|  | **Von Christiane Wiemann \\| frauwiemann.de** | \n|  | Nutze diese zwei Prompts nacheinander, um aus einem wertvollen ChatGPT-Gespräch | \n|  | eine saubere Markdown-Handover-Datei zu erstellen – bereit für ein neues Projekt | \n|  | oder eine andere KI-Umgebung. | \n|  | --- | \n|  | ## Schritt 1: Handover erstellen | \n|  | Act as a senior project archivist, knowledge manager, and AI handover editor. | \n|  | Your task is to turn this entire conversation into a high-quality Markdown handover document that can be saved and uploaded into a future project, workspace, or AI system as compact long-term context. | \n|  | The goal is not to create a generic summary. The goal is to preserve the most valuable knowledge from this conversation so that a future AI assistant can continue the work with minimal loss of context, decisions, tone, and strategic intent. | \n|  | Please analyze the full conversation carefully. | \n|  | Create a Markdown document with the following sections: | \n|  | # PROJECT HANDOVER SUMMARY | \n|  | ## 1. Project Purpose | \n|  | Summarize what this project is about, the overall objective, the role the AI has played, and what success looks like. | \n|  | ## 2. Current Status | \n|  | Capture where the project stands, what has been completed, what is still open, and what is in progress. | \n|  | ## 3. Important Milestones | \n|  | List the most relevant milestones chronologically. For each: what happened, why it mattered, what changed afterwards. | \n|  | ## 4. Key Decisions | \n|  | Extract all meaningful decisions. For each: Decision, Rationale, Implication, Status (confirmed / provisional / still open). | \n|  | ## 5. User Preferences and Working Style | \n|  | Capture preferences that materially influence future work: tone, writing style, formatting, level of detail, decision-making style, things the user dislikes. | \n|  | ## 6. Feedback and Learning | \n|  | Summarize what worked, what did not, corrections made, lessons from previous iterations. | \n|  | ## 7. Important Context | \n|  | Capture relevant context: stakeholder relationships, organizational context, constraints, dependencies, tools, commercial considerations. | \n|  | ## 8. Tone, Language and Communication Style | \n|  | Describe the desired communication style: preferred tone, vocabulary, sentence rhythm, formality, use of humour or directness. | \n|  | ## 9. Explicit Rules and Guardrails | \n|  | List rules that have emerged: what should always be done, what should never be done, what requires verification. | \n|  | ## 10. Open Questions | \n|  | List unresolved questions, decisions, uncertainties or dependencies. Do not turn them into assumptions. | \n|  | ## 11. Next Best Actions | \n|  | Suggest the most sensible next steps. Separate agreed next steps from AI-recommended ones. | \n|  | ## 12. Hidden but Important Patterns | \n|  | Review for patterns not always stated directly but supported by repeated behaviour. Label as \"Inferred pattern\". | \n|  | ## 13. What a Future AI Should Remember | \n|  | The 10–20 most important facts, rules, preferences and decisions. Practical and easy to scan. | \n|  | ## 14. Source Integrity | \n|  | Three lists: Explicitly stated / Strongly inferred / Uncertain – should be verified. | \n|  | OUTPUT REQUIREMENTS: | \n|  | - Clear, structured Markdown | \n|  | - Compact but information-dense | \n|  | - No repeated information across sections | \n|  | - Preserve names, terminology, dates, decisions | \n|  | - Do not invent facts | \n|  | - Flag contradictions rather than silently correcting them | \n|  | - Remove filler and casual conversation | \n|  | - Preserve strategically important nuance | \n|  | - Write so that another AI could continue the work immediately | \n|  | At the end, propose a filename: PROJECTNAME_Handover_YYYY-MM-DD.md | \n|  | --- | \n|  | ## Schritt 2: Qualitätscheck (direkt danach senden) | \n|  | Review the handover summary you just created. | \n|  | Check whether any important decision, correction, stakeholder nuance, recurring user preference, project constraint, or unresolved issue from the original conversation is missing. | \n|  | If something relevant is missing, add it. | \n|  | If anything in the summary is speculative or too confidently phrased, mark it clearly as inferred or uncertain. | \n|  | Then return the final cleaned Markdown version. | \n|  | --- | \n|  | *Schritt 2 ist entscheidend. Er holt genau das heraus, was zwischen den Zeilen wichtig geworden ist.* |", "url": "https://wpnews.pro/news/gpt-handover-prompt-md", "canonical_source": "https://gist.github.com/JaneWee8/b8e6defd009ce34e3f107b7919a03451", "published_at": "2026-10-04 14:59:08+00:00", "updated_at": "2026-10-04 15:12:13.347097+00:00", "lang": "en", "topics": ["ai-tools", "large-language-models", "generative-ai"], "entities": ["Christiane Wiemann", "frauwiemann.de", "ChatGPT"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/gpt-handover-prompt-md", "markdown": "https://wpnews.pro/news/gpt-handover-prompt-md.md", "text": "https://wpnews.pro/news/gpt-handover-prompt-md.txt", "jsonld": "https://wpnews.pro/news/gpt-handover-prompt-md.jsonld"}}