cd /news/ai-tools/gpt-handover-prompt-md · home › topics › ai-tools › article
[ARTICLE · art-144876] src=gist.github.com ↗ pub= topic=ai-tools verified=true sentiment=· neutral

GPT_Handover_Prompt.md

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.

by read4 min views1 publishedOct 4, 2026

| | # GPT Handover Prompt – Two-Step Version | | | Von Christiane Wiemann | frauwiemann.de | | | Nutze diese zwei Prompts nacheinander, um aus einem wertvollen ChatGPT-Gespräch | | | eine saubere Markdown-Handover-Datei zu erstellen – bereit für ein neues Projekt | | | oder eine andere KI-Umgebung. | | | --- | | | ## Schritt 1: Handover erstellen | | | Act as a senior project archivist, knowledge manager, and AI handover editor. | | | 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. | | | 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. | | | Please analyze the full conversation carefully. | | | Create a Markdown document with the following sections: | | | # PROJECT HANDOVER SUMMARY | | | ## 1. Project Purpose | | | Summarize what this project is about, the overall objective, the role the AI has played, and what success looks like. | | | ## 2. Current Status | | | Capture where the project stands, what has been completed, what is still open, and what is in progress. | | | ## 3. Important Milestones | | | List the most relevant milestones chronologically. For each: what happened, why it mattered, what changed afterwards. | | | ## 4. Key Decisions | | | Extract all meaningful decisions. For each: Decision, Rationale, Implication, Status (confirmed / provisional / still open). | | | ## 5. User Preferences and Working Style | | | Capture preferences that materially influence future work: tone, writing style, formatting, level of detail, decision-making style, things the user dislikes. | | | ## 6. Feedback and Learning | | | Summarize what worked, what did not, corrections made, lessons from previous iterations. | | | ## 7. Important Context | | | Capture relevant context: stakeholder relationships, organizational context, constraints, dependencies, tools, commercial considerations. | | | ## 8. Tone, Language and Communication Style | | | Describe the desired communication style: preferred tone, vocabulary, sentence rhythm, formality, use of humour or directness. | | | ## 9. Explicit Rules and Guardrails | | | List rules that have emerged: what should always be done, what should never be done, what requires verification. | | | ## 10. Open Questions | | | List unresolved questions, decisions, uncertainties or dependencies. Do not turn them into assumptions. | | | ## 11. Next Best Actions | | | Suggest the most sensible next steps. Separate agreed next steps from AI-recommended ones. | | | ## 12. Hidden but Important Patterns | | | Review for patterns not always stated directly but supported by repeated behaviour. Label as "Inferred pattern". | | | ## 13. What a Future AI Should Remember | | | The 10–20 most important facts, rules, preferences and decisions. Practical and easy to scan. | | | ## 14. Source Integrity | | | Three lists: Explicitly stated / Strongly inferred / Uncertain – should be verified. | | | OUTPUT REQUIREMENTS: | | | - Clear, structured Markdown | | | - Compact but information-dense | | | - No repeated information across sections | | | - Preserve names, terminology, dates, decisions | | | - Do not invent facts | | | - Flag contradictions rather than silently correcting them | | | - Remove filler and casual conversation | | | - Preserve strategically important nuance | | | - Write so that another AI could continue the work immediately | | | At the end, propose a filename: PROJECTNAME_Handover_YYYY-MM-DD.md |

|  | --- | 
|  | ## Schritt 2: Qualitätscheck (direkt danach senden) | 

| | Review the handover summary you just created. | | | Check whether any important decision, correction, stakeholder nuance, recurring user preference, project constraint, or unresolved issue from the original conversation is missing. | | | If something relevant is missing, add it. | | | If anything in the summary is speculative or too confidently phrased, mark it clearly as inferred or uncertain. | | | Then return the final cleaned Markdown version. | | | --- | | | Schritt 2 ist entscheidend. Er holt genau das heraus, was zwischen den Zeilen wichtig geworden ist. |

── more in #ai-tools 4 stories · sorted by recency
── more on @christiane wiemann 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
→ Live at https://your-agent.zahid.host ✓
Get free account → Pricing
from €0/mo · no card required
LIVE [news/gpt-handover-prompt-…] indexed:0 read:4min 2026-10-04 · —