test-applisted

Exploratory testing of the application itself — what `test-tests` does to a test suite, this does to the running product. Launch the workflow; do **not** drive the app inline. The workflow returns a structured report; relay it and save it to a file. **This runs the application for real.** Every iteration launches it against whatever environment it is configured for and uses it as a user would — creating, changing, deleting and sending real data. The skill applies no safety constraints and cannot: it has no way to know what the app reaches. Point it at an environment whose data you can afford to lose. ## Run the workflow 1. Parse `$ARGUMENTS` as `[low|medium|high|ultra] [focus-or-doc]`. Both optional. A leading `low` | `medium` | `high` | `ultra` token is the **level**; everything after it is the **focus**, in one of three shapes: - **absent** — launch the application and exercise it broadly. - **an area** ("the export feature") — exercise that part of it. - **instructions, or a path to a doc** — do that specifically. A doc is both the scope and the yardstick: every claim in it becomes an expectation, and the app disagreeing with it is a finding. If no level token is given, ask with `AskUserQuestion` (header "Level"): - `low` — 2 iterations, documented happy paths only. The rung to pick when you are unsure what the environment can absorb. - `medium` (recommended) — 4 iterations, plus edge cases and unhappy paths. - `high` — 8 iter