The prompt is a sideways way to claim a strength — what others actually trust you with reveals more than self-claimed virtues. Strong answers name one calibrated domain; weak ones reach for résumé bullets or universal advice categories.
...how to write a tough text. The 'thinking of you, not asking for anything' kind. I'm the friend who edits.
Why it works: Specific narrow domain ('the tough text'), with a specific subgenre and a quoted line. The 'I'm the friend who edits' close lands as identity, not flex.
...what to cook when someone you love has had a rough week. I keep a list.
Why it works: Names a real situation, then a tiny operational detail ('I keep a list') that proves it isn't theoretical. Signals practical care without naming care as a virtue.
...whether to say the thing or not say the thing. I'm usually team say it, calmly.
Why it works: Tiny domain with a clear personal verdict. The 'calmly' is the calibration — signals the answerer has thought about how, not just whether.
...career stuff. People come to me for negotiation tips.
Why it falls flat: Work skill in the wrong place. Puts the matcher in job-interview mode and signals the answerer's identity is mostly their professional reputation.
...life. Honestly, all of it.
Why it falls flat: Refuses to commit to one domain. Universal vague claim with no observable signal — the matcher gets nothing concrete to picture.
...how to make better decisions in general.
Why it falls flat: Self-help abstraction. Sounds like a wise position but names no real situation, no real expertise, no observable behavior. Filters no one.
The prompt rewards a narrow real domain that signals a specific kind of attention. The strongest answers name one calibrated thing (the tough text, the meal for someone having a rough week, the say-it-or-not call) with a tiny operational detail that proves it's real. The most common failure is the work-flex ('career stuff, negotiation tips') which puts the matcher in interview mode. The second is the vague gesture ('life, all of it') which refuses to commit. The third is the self-help abstraction ('better decisions in general') which sounds wise and names nothing observable. Pick the small narrow domain and prove it with a detail.
Pick one narrow specific domain with a tiny operational detail (the list you keep, the verdict you usually give, the kind of message you edit). Narrow-and-real beats broad-and-impressive — the matcher gets a precise read on what you pay attention to.
Usually no. Career advice domains ('comp negotiation', 'startup strategy') put the matcher in job-interview mode and signal that the answerer's social currency is mostly their work reputation. Save those for LinkedIn — pick something interpersonal here.
Because the prompt's whole job is to surface one specific kind of attention you offer. Universal claims ('I help with everything') refuse the prompt's narrowing function and give the matcher nothing concrete. The fix is to commit to one tiny domain and prove it.
Once your prompts land, the next bottleneck is the messages. Opening lines tuned to her bio, replies that actually land, and a free profile roast.
Try the opening-lines tool freeOne tap with Google. No card.