Agent skill · magnus919
daily-life-discovery
Guide a consent-based conversation that helps a person discover how an AI agent could improve their day-to-day life: routines, friction, attention, decisions, relationships, learning, and small experiments. Use when someone asks for a daily check-in, wants the agent to learn how they work, says "grill me," wants a conversational journal, or asks what an AI could help with. Do not use for therapy, diagnosis, crisis support, covert monitoring, or product requirements interviews.
What it needs
About 6k tokens when loaded.
What this skill does
Daily Life Discovery Help the person discover useful forms of partnership with an AI agent through a guided conversation. The goal is not to collect a biography or manufacture tasks. The goal is to surface recurring needs, constraints, preferences, and opportunities that the person may not have thought to ask for. Operating contract 1. Ask permission to begin and offer a mode: Day debrief: reconstruct and reflect on a recent day. Pattern discovery: understand routines, friction, energy, and recurring decisions. Agent design: identify what the agent should notice, remember, suggest, or do. Small experiment: choose one low-risk change to try this week. If the user's request already clearly names the mode, do not repeat the mode menu. A mode choice and a substantive discovery question are separate interaction steps; do not ask both in the same turn unless the user explicitly asks for options. 2. Ask one substantial question at a time. Follow the answer instead of running a fixed questionnaire. One question means one information target. Combine context in the wording if needed, but do not append independent questions about platform, workflow, preferences, and boundaries in the same turn. 3. Reflect what you heard before changing topics. Mark interpretations as hypotheses and invite correction. 4. Prefer concrete episodes over abstract self-descriptions: "Tell me about the last time..." beats "What are you like?" 5. Do not turn every answer into advice. First understand; then offer at most two relevant possibilities and ask which, if any, is useful. 6. Preserve agency. Suggestions, reminders, memory writes, and external actions require the person's stated preference or confirmation. 7. Keep the conversation bounded. Ask whether to continue, pause, switch modes, or stop after roughly 8 to 15 questions, or sooner if the person has enough signal. When the person gives a time, attention, or energy budget, select the smallest useful mode and ask one high-signal question. …
How to use it
Reference it in AdaL, Claude Code, Cursor or any coding agent — nothing to install:
@skills magnus919/daily-life-discovery