# Website Audit & Client Handoff Kit

Public preview edition · November 12, 2026 · Not for sale

## What you are opening

Twenty practical website checks, seven editable files and one fictional worked finding. This is a process for turning observations into owned, testable work. It is not an automated audit, legal opinion, security assessment, accessibility certification or revenue forecast. This usable preview is not for sale. See https://www.mortenagiraffe.com/terms for the current resource terms.

## Start with a decision

1. Write the visitor, their main task and the intended next step. Name the project owner and release approver.
2. Agree the environment and what may be tested. Default to local/mocked services; do not submit forms, book appointments, send emails, change providers or deploy without explicit approval.
3. Open check-register.csv in your spreadsheet. Pick the relevant checks. Every row starts **Not assessed**. Filtering a list is not performing an audit.
4. Reproduce one concrete failure. Save the affected route, state, browser/device and exact observation in evidence-log.csv. Do not place customer data, credentials or tokens in the evidence.
5. Assign an owner and acceptance test. Use scope-worksheet.md to identify dependencies and estimate only after understanding the change.
6. Fix the cause, repeat the relevant test and record the result. A plausible fix or passing build alone does not establish a working user flow.
7. Use client-report.md to explain priorities and decisions. Use handoff.md only when the release and verification states are known.

## Priority rubric

- P0: verified blocker of a critical task or serious data/security/privacy failure; stop the affected release and involve the responsible owner.
- P1: material visitor friction or misleading task/content behavior; address before lower-impact polish.
- P2: refinement or non-blocking issue with a documented benefit.
- Untriaged: evidence or context is insufficient. Do not invent precision.

Severity is not a claim of lost sales. Add reach, evidence confidence, effort and dependencies in the decision notes. A low-effort fix is not automatically the highest-impact fix.

## Status vocabulary

- Not assessed: no check performed.
- Open: a supported finding needs action.
- Needs decision: an owner must resolve scope, policy, rights or another dependency.
- Ready to verify: the change exists, its acceptance test has not passed yet.
- Failed: the acceptance test did not pass.
- Verified: named test passed on the recorded source/environment.
- Deferred: named owner accepted a gap with a revisit trigger.
- Not applicable: include a reason; it is not a pass.

## File map

| File | Job |
|---|---|
| check-register.csv | 20 checks, manual tests, evidence and acceptance criteria |
| evidence-log.csv | A reproducible observation and verification trail |
| client-report.md | A decision-focused summary, with fictional worked finding |
| scope-worksheet.md | Ownership, dependencies, estimates and change control |
| handoff.md | Release receipt, remaining gaps, rollback and ownership |
| prompt-pack.md | 20 bounded prompts, one for each check |
| quick-start.md | Workflow, limits and status definitions |

Your existing Morten Journal, tools, Design Blueprint and launch downloads remain free. This proposed product’s extra value is the connected editable workflow and deliverables, not exclusive access to a checklist.
