← ClaudeAtlas

quickstart-verifylisted

Execute a README / getting-started / CONTRIBUTING setup section end-to-end in a CLEAN environment to prove onboarding actually works — install → build → run → smoke. Reports per-step pass/fail and time-to-first-green. Distinct from doc-writer, which authors the prose; this one runs it.
adnanmokhtar/refract · ★ 1 · AI & Automation · score 80
Install: claude install-skill adnanmokhtar/refract
# quickstart-verify ## Premise The failure mode: a getting-started section that *looks* right and doesn't *run*. A missing prereq the maintainer forgot they'd installed years ago, a command renamed two releases back, an env var nobody documented, steps in the wrong order. This is the #1 cause of onboarding abandonment — the new contributor hits step 4, it errors, and they leave. Doc-drift-scan catches docs that reference dead *code*; this catches docs that reference a *procedure* that no longer produces a working checkout. `smoke-verify` made the same leap for code: a green test suite does not prove the app boots, so it stops trusting the tests and actually starts the process. quickstart-verify makes that leap for onboarding docs: a setup section that reads correctly does not prove a human can follow it, so it stops trusting the prose and actually executes every step in a clean environment. Cite-or-halt. Every failing step cites **the doc line + the exact command run + the actual error output**. "The setup seems off" is a vibe, not a finding. A failure without the captured command and stderr is not reportable. Boundary: **doc-writer WRITES the prose; quickstart-verify EXECUTES and PROVES it.** This skill does not author onboarding narrative, restructure the README, or improve wording. It runs the documented steps, and where a step fails it emits the minimal *setup* patch (add the missing prereq, fix the stale command, correct the order). Anything past the setup steps is