market-competitor-research

Run round-based parallel web searches: first map the broad landscape and named competitors, then drill into differentiators, then fill coverage gaps. Seed the research from a curated awesome-list on GitHub to map dozens of projects quickly, then verify key entries before citing them.

Clone the repo with git clone --depth 1 and grep the source for the README's headline claims, checking feature keywords, test files, and module names. Report a docs-vs-implementation table and flag any repo where claims exceed the actual code.

A decision matrix scores each candidate 1-5 on Pain, Fit, Coverage, First-party risk, and Defensibility, then assigns a BUILD, WATCH, or DROP verdict with a one-clause rationale. It ends with a positioning section naming what to build and what not to compete with.

Use RDAP: query https://rdap.org/domain/<name> where a 404 means registry-confirmed unregistered and a 200 means registered. An empty dig A record only means no DNS record, and registrar widgets are often bot-walled, so RDAP is the authoritative source.

Web extraction tools return 'Website Not Supported' errors on Reddit. Instead, use web search to surface Hacker News threads, which include full comment text and carry the real signal for developer-tool sentiment.

Fast-growing repos invalidate third-party summaries within months, since a repo can gain thousands of stars and full test suites after a claim was made. When a community claim contradicts the repo's code, trust the cloned code.