← ClaudeAtlas

verify-melisted

Check that a built feature actually matches the original problem (/grill-me) and the chosen solution (/solve-me). Acts like a real user and a QA engineer, not a code reviewer: confirms scope, environment, database safety and login with the developer first, then creates real dummy data in the database (via Laravel MCP/Boost tools if available, otherwise artisan), then hits the real running app with curl — golden path, edge cases, and heavy/weird inputs. No automated tests are written. Produces one clean markdown file with a simple, readable log of every request tested (input/output) plus a plain-language list of the unit/feature tests that still need to be written, and then stops — it never writes those tests and never starts another skill. Use after /build-me, once a commit (or the whole feature) is implemented.
Kerliula/ship-me · ★ 1 · AI & Automation · score 72
Install: claude install-skill Kerliula/ship-me
# Verify Me — Prove It Actually Works Your job is to find out whether the built feature really does what the problem and solution said it should — by using it, the way a real user or an outside API caller would. You are not reading the code to judge it. You are poking the running app and writing down what happens. You do not write unit tests, feature tests, or any other code. You do not fix bugs. You observe, compare against what was promised, and report — cleanly. --- ## Input You need three things: 1. The problem write-up (from `/grill-me`, e.g. `docs/grilling/<topic>.md`). 2. The solution write-up (from `/solve-me`, e.g. `docs/solutions/<topic>.md`). 3. What was actually built (from `/build-me` — recent commits, changed files, or just ask the developer what to verify). If any of these are missing, ask for them or the file paths. Don't guess at what "correct" means — pull it from the problem's rules and edge cases and the solution's chosen options. ### Confirm with the developer before touching anything Even when all three inputs are handed to you, **stop once, ask, and wait** before running the first request. Keep it to one short round — propose your own answers so they can just say "yes": 1. **What am I verifying?** — the list of R-numbers / behaviors you intend to hit, and anything you're deliberately skipping. 2. **Where do I run it?** — which base URL / environment, and is the server already running or should you start it. 3. **Is this database