Release decision report
Fictional sample. Every metric, status, organization, and decision on this page exists only to demonstrate the report engine; replace it with verified project evidence before use.
This starter organizes a real review: the decision, the evidence behind it, the remaining risk, and the next accountable steps. Replace the sample facts while keeping the semantic structure.
Decision signal
Verified findings
| Signal | Evidence | State |
|---|---|---|
| Authoring contract | Registry and schema checks | Confirmed |
| Portable output | Desktop and mobile file:// journeys |
Confirmed |
| Package boundary | Clean-consumer installation | Confirmed |
| External release | Explicitly outside this report | Not started |
The table separates observed facts from work that has not happened. Keep that distinction when replacing the sample content.
Evidence map
Scope
The candidate includes the declarative authoring loop, static output, and package-owned interactions.
Confidence
Focused unit, browser, and installed-package journeys cover the public path.
Boundary
Publication and deployment remain separate external actions.
Decision and review
Accept the candidate for release preparation
The evidence supports advancing. Any new blocking observation reopens this decision with a reproducible failure, owner, and next check.
State the decision
Record the audience, scope, and success criteria in ordinary Markdown.
Replace the sample
Keep the structure and replace every demonstration claim with observed evidence.
Create the artifact
Generate the portable page in the selected output format.
Review the result
Open the built file and record the decision against observed evidence.
Review detail
Open the residual-risk register
- Large embedded assets can exceed the configured warning threshold.
- External publication still requires an explicit release action.
- New requirements need their own evidence before they enter this decision.
Next actions
Complete the handoff
- Replace the sample findings with verified project facts.
- Build the selected output;
buildvalidates the source before writing it. - Open the result directly in a browser and review the real page.
- Use
validateorinspectonly when you need focused diagnostics, then assign owners and dates.
Review confidence
Use this bounded control during a live review to count independently confirmed evidence groups.
