Quick debrief after a date, trip, or activity. Four questions, under 2 minutes, auto-updates memory with sentiment tags and seasonal data.
/feedback → Start a debrief
/feedback hiking date last weekend → Skip Q1, jump to rating
What you get: Memory files updated with rating, learning, sentiment tags, and season data. Future suggestions reflect what worked, why it worked, and what time of year works best.
Time: Under 2 minutes.
If an argument was provided (e.g., "hiking date last weekend"), use it as the answer to Q1 and skip straight to Q2.
Otherwise, ask exactly these four questions in order. Accept short answers.
Q1: "What was it? (the date, trip, or activity — name or describe it briefly)"
Q2: "How did it go?" Accept natural language. Map to rating internally:
Q3: "What made it work (or not)? Pick any that apply, or say it in your own words:"
Tags: new-discovery | intimate | energising | relaxing | creative | social |
adventurous | familiar-comfort | food-highlight | too-crowded | too-expensive |
wrong-energy | bad-timing
Accept their own words too — map to the closest tag(s) internally. Multiple tags are fine and encouraged.
Q4: "One thing worth remembering for next time?" This is the learning. Accept anything: "venue was too loud," "the sunset timing was perfect," "skip the appetizers," "do this again in winter."
After collecting answers, Claude determines these tags automatically — do NOT ask:
Season tag: Based on the date of the activity:
summermonsoonautumnwinterspringSetting tag: Based on the activity description:
indoor or outdoor or mixedCategory tag: Based on the activity type:
food | adventure | cultural | active | relaxation | home | travel | class | socialUpdate the right files:
If it was a date or activity: Add entry to context-library/memory/past-dates.md
Format: | [YYYY-MM-DD] | [Title] | [❤️/👍/👎] | [sentiment tags] | [season:setting] | [Note from Q4] |
If it was a trip: Add entry to context-library/memory/past-trips.md
Format: | [YYYY-MM] | [Destination] | [Duration] | [❤️/👍/👎] | [sentiment tags] | [Note] |
If Q4 contains a preference signal: Also add to context-library/memory/learnings.md
under ## General Preferences:
Format: - [YYYY-MM-DD]: [Observation]
Update Sentiment Patterns in learnings.md:
Under ## Sentiment Patterns, update the tag count table:
Update Seasonal Patterns in learnings.md:
Under ## Seasonal Patterns, update the season row:
Update suggestion-log.md: If the date came from a system suggestion (check outputs/suggestion-log.md), mark it as "done + [rating]" in the log.
Write to Notion + local (except /surprise dates — those are local only until the couple explicitly logs them via /feedback)
Got it. Updated [files changed]:
→ [what was logged, one line]
→ Tagged: [sentiment tags] | [season:setting]
[If a learning was extracted:] Also noted: "[learning]"
[If a sentiment pattern emerged:] Pattern update: your best dates tend to be [dominant tags].
No follow-up questions. Keep it fast. Done.