# Calibrated example reviews

These examples demonstrate the decision logic; they are not actual funding decisions.

## Example 1 — Accept: reproducible accessibility baseline

- Contribution: `example-accessibility-baseline`
- Opportunity version: `2026-08-01T10:00:00Z`
- Artifacts checked: public report, rerun instructions, verifier, release manifest

### Objective gate

| Gate | Result | Evidence / observation |
|---|---|---|
| G1 | PASS | Six required pages are mapped to six report sections. |
| G2 | PASS | Report and evidence URLs returned successfully at review time. |
| G3 | PASS | Every named page has findings, priority, and a next action. |
| G4 | PASS | Primary sources and original observations are separated and linked. |
| G5 | PASS | Twenty deterministic checks pass; manifest records artifact hashes. |
| G6 | PASS | Review uses public pages only and requires no account or sensitive action. |
| G7 | PASS | README describes inputs, rerun command, output, and limitations. |

Gate decision: **ACCEPTABLE**.

### Quality judgment

| Dimension | Score | Reason |
|---|---:|---|
| Q1 Usefulness | 4 | Prioritized findings translate directly into an improvement backlog. |
| Q2 Clarity | 3 | Navigation is strong; one summary table could be shorter. |
| Q3 Reproducibility | 4 | Inputs, commands, expected results, and hashes are present. |
| Q4 Rigor | 3 | Dynamic-state limitations are disclosed; assisted-technology testing remains future work. |
| Q5 Maintainability | 3 | Versioned manifest and stable selectors support reruns. |

Weighted score: **3.45 / 4.00 (strong, advisory)**.

Recommendation: **Accept**. Optional improvement: add a compact change log to the next release.

Grant eligibility: **Eligible for discretionary consideration** because all gates pass and the work is reusable. Eligibility is not an award, payment promise, ranking, or guarantee. Any grant requires a separate authorized decision and available project funds.

## Example 2 — Request changes: documentation package with broken evidence

- Contribution: `example-integration-guide`
- Opportunity version: `2026-08-02T12:00:00Z`
- Artifacts checked: guide URL, claimed source archive, release note

### Objective gate

| Gate | Result | Evidence / observation |
|---|---|---|
| G1 | PASS | The guide addresses installation, configuration, and rollback. |
| G2 | FAIL | The source-archive URL returns 404. |
| G3 | NEEDS EVIDENCE | The rollback procedure is described but not demonstrated. |
| G4 | NEEDS EVIDENCE | Two copied diagrams have no source or license note. |
| G5 | FAIL | The release note lists a hash, but there is no downloadable artifact to verify. |
| G6 | PASS | No unsafe or restricted action is required. |
| G7 | FAIL | A contributor cannot reproduce the documented example without sample configuration. |

Gate decision: **CHANGES REQUIRED**.

### Quality judgment

| Dimension | Score | Reason |
|---|---:|---|
| Q1 Usefulness | 2 | Core flow is useful, but missing inputs prevent execution. |
| Q2 Clarity | 3 | Instructions are readable and ordered. |
| Q3 Reproducibility | 1 | Broken archive and absent sample configuration block a rerun. |
| Q4 Rigor | 2 | Rollback claim and diagram provenance need evidence. |
| Q5 Maintainability | 2 | Versioning exists, but release assets are incomplete. |

Weighted score: **2.00 / 4.00 (adequate boundary, advisory)**.

Recommendation: **Request changes**. Required: restore the source archive, publish the sample configuration, verify the rollback example, disclose diagram provenance/licenses, and provide hashes for the restored artifacts. Re-review only G2–G5 and G7 if the opportunity version is unchanged.

Grant eligibility: **Not currently eligible** because objective gates fail. Eligibility is not an award, payment promise, ranking, or guarantee. Any grant requires a separate authorized decision and available project funds.

## Example 3 — Accept with caveats: public dataset with bounded uncertainty

- Contribution: `example-community-index`
- Opportunity version: `2026-08-04T09:30:00Z`
- Artifacts checked: CSV, data dictionary, methodology, validation output

### Objective gate

| Gate | Result | Evidence / observation |
|---|---|---|
| G1 | PASS | CSV and methodology cover all requested entities and fields. |
| G2 | PASS | Files download without authentication. |
| G3 | PASS | Required columns are present; nullable fields are documented. |
| G4 | PASS | Every row includes a public source URL and retrieval date. |
| G5 | PASS | Schema validation passes; duplicate and null counts are published. |
| G6 | PASS | Dataset contains public organizational data, not sensitive personal data. |
| G7 | PASS | Data dictionary, refresh steps, and known limitations are included. |

Gate decision: **ACCEPTABLE**.

### Quality judgment

| Dimension | Score | Reason |
|---|---:|---|
| Q1 Usefulness | 3 | Immediately usable for discovery and analysis. |
| Q2 Clarity | 4 | Field definitions and null semantics are explicit. |
| Q3 Reproducibility | 3 | Validation is repeatable; collection still needs some manual review. |
| Q4 Rigor | 4 | Stale and contradictory sources are flagged rather than silently resolved. |
| Q5 Maintainability | 2 | Refresh ownership and cadence are suggestions, not commitments. |

Weighted score: **3.30 / 4.00 (strong, advisory)**.

Recommendation: **Accept**, with caveats. Required changes: none. Optional: assign refresh ownership and add an automated freshness warning. Limitation: acceptance confirms the documented snapshot, not continued accuracy of external sources.

Grant eligibility: **Eligible for limited discretionary consideration** based on utility and rigor; the non-binding amount or band must be decided separately by an authorized project steward. Eligibility is not an award, payment promise, ranking, or guarantee. Any grant requires a separate authorized decision and available project funds.