Every engagement ends with one of these. Not a meeting. Not a vague "handoff." A receipt — with evidence, metrics, and a checklist you can hold us to.
The sample below is for a real-type engagement: QA automation for a 4-person Shopify shop. Annotated callouts explain what each section does and why it matters.
A machine-readable, human-readable document that replaces the "can you remind me what was included?" email six months later. Committed to GitHub with the project.
Scope creep happens. This locks what was agreed vs. what shipped — protects both sides.
Any freelancer can say "tests are passing." We show you the run. Click the link — it's real. This is the differentiator vs. Toptal and offshore QA.
Vague deliverables get disputed. Numbers don't. We baseline before we start so the delta is provable at the end.
| Metric | Before | After | Delta |
|---|---|---|---|
| Playwright test coverage (critical paths) | 0% | 100% | +100% |
| Flaky test rate | 22% | 3% | −87% |
| Time to catch checkout regressions | Manual (2–3 days) | Automated (<5 min) | 99% faster |
| CI pipeline pass rate (last 30 runs) | N/A (no CI) | 96% | New baseline |
Most freelancers leave. We hand off. The difference is this list — every operational item checked before the engagement closes.
docs/qa-runbook.md in main repo — covers failure triage, selector updates, and adding new tests
docs/qa-runbook.md § badgesNot legal theater — it's a shared moment of closure. Both sides agree: this is done. No lingering ambiguity.
Every engagement starts with a defined scope and ends with a receipt. No ambiguous handoffs. No "it was working on my machine."
Book an engagement →