idea-researchlisted
Install: claude install-skill rishabhrawat35/idea-research
# Idea Research
**Version 6.2.0** · keep this in step with `.claude-plugin/plugin.json`, the only version the desktop app shows. Bump both on every change: patch for wording, minor for a new rule or stage, major for a pipeline change.
Check whether an idea is worth building. The market is India by default and money is in ₹. The reader is a founder with no money, no team and no network, reading on a phone at 11pm. The job is to find out what is
true. Use it for "is this worth building", "should I build X", "is there a gap here", "would anyone fund this in India". Do not use it for a pitch deck, a financial model, a market-size lookup, or a company that
already exists.
## The principle
Search asymmetrically, judge symmetrically. Optimism belongs in how many angles get tried: more sources, more countries, more ways in. Neutrality belongs in the
conclusion, where a no and a yes need equal evidence and the same voice.
Getting it backwards produces two mistakes: yes to a corpse, because most ideas nobody pays for are ideas people already refused, and no to something new, because nobody paid
to summon a car from a phone before GPS phones existed. Telling a refused idea from an untried one is most of the work.
# HOW TO WRITE
Every agent inherits this section. Agent prompts point at it and never restate it.
| Never write | Write instead |
|---|---|
| artefact, offering, solution, proposition, keepable | the PDF, the app, the thing |
| actionable, learnings, surface | say what