This prompt goes beyond summarizing a chat transcript. It extracts decisions, reusable lessons, friction, and unfinished work for an end-of-day or milestone review.
It may use only records that are actually accessible. If the environment cannot see the complete history, the output must state its coverage instead of implying that every conversation was reviewed.
Copy-ready prompt
Review the conversations that are actually accessible for the specified date or time range. Produce a concise, technical, action-oriented work reflection.
Start by stating:
- which conversations or time range were reviewed
- which history was unavailable
- any material evidence gaps
Treat paths, commits, validation results, completion states, and decisions as facts only when they appear in the accessible record. Do not infer unseen work. Consolidate repeated discussions into the latest supported conclusion, and distinguish explicit decisions from patterns you infer across conversations.
Return:
1. Core direction
- the real goals behind the conversations
2. Progress and decisions
- completed work, confirmed trade-offs, and verified results
3. Reusable lessons
- rules, judgments, or working methods worth preserving
4. Friction and noise
- repetition, scope drift, failed approaches, and avoidable back-and-forth
5. Open loops
- unfinished work, unresolved choices, and explicit blockers
6. Next actions
- the top 1–3 concrete, high-leverage actions
7. Long-term signal
- one clearly labeled observation, without presenting inference as fact
When a section lacks evidence, write No supported item. Avoid chronological retelling, generic encouragement, and invented productivity judgments.Practical use
For a daily review, provide the date, timezone, and accessible conversation scope. For a milestone review, include related task, document, or Git evidence so completion is not inferred from conversational wording alone.