softhouse-uatlisted

Runs the current project's UAT acceptance suites against its live or staging system, resolving the UAT definition from .softhouse/uat.md, a CLAUDE.md verification section, or a generic fallback. Use when the user runs /softhouse-uat, asks to run UAT or acceptance tests, or needs to verify a deployed system end-to-end.
buya-v/claude-softhouse · ★ 0 · AI & Automation · score 70

Install: claude install-skill buya-v/claude-softhouse

# /softhouse-uat — Run UAT Acceptance Tests (v2, project-parameterized) Runs the CURRENT project's UAT suite against its live/staging system. v2: this skill no longer hardcodes any project — the UAT definition is resolved from the repo you are in. ## Usage - `/softhouse-uat` — Run all UAT suites for this project - `/softhouse-uat <suite>` — Run one named suite (names come from the project's UAT config) ## Resolution order (STEP 0) Resolve the UAT definition for the current working directory, in this order: 1. **`.softhouse/uat.md`** in the repo — the canonical per-project UAT config. It declares: - **Pre-flight**: health/smoke URLs or commands that must succeed before testing - **Deploy/refresh**: how to get the system under test current (and whether that is orchestrator-only / user-scheduled — NEVER deploy a live system implicitly) - **Suites**: named suites with their exact commands - **Pass criteria** and where evidence (reports/screenshots) lands 2. A **"UAT" or "Verification" section** in CLAUDE.md or the project's OPERATIONS doc. 3. **Generic fallback** (no config found): - (a) full test suite green in an ISOLATED stack (never the live default compose project) - (b) live smoke: health endpoint returns 200 + one end-to-end domain action succeeds (e.g. an authenticated API round-trip with a test credential) - (c) for UI changes: fetch the changed pages, assert HTTP 200 and no raw errors ("undefined", "NaN", stack traces) in the HTML - Report w