cms-field-vocabulary 1.0.11 → 1.2.0

package/CHANGELOG.md

@@ -1,3 +1,15 @@
+## [1.2.0](https://github.com/JFusco/cms-field-vocabulary/compare/v1.1.0...v1.2.0) (2026-09-26)
+
+### Features
+
+* **sources:** complete SitecoreAI authoring list fields ([14d11ee](https://github.com/JFusco/cms-field-vocabulary/commit/14d11ee693eb177cf1ea5b6f528a34308bdb1ce3))
+
+## [1.1.0](https://github.com/JFusco/cms-field-vocabulary/compare/v1.0.11...v1.1.0) (2026-09-26)
+
+### Features
+
+* **sources:** add Optimizely CMS SDK 3.0 profile ([974aee6](https://github.com/JFusco/cms-field-vocabulary/commit/974aee6199f80fbcbd3bfbd4d6eae3c02e6995e3))
+
 ## [1.0.11](https://github.com/JFusco/cms-field-vocabulary/compare/v1.0.10...v1.0.11) (2026-09-26)
 
 ## [1.0.10](https://github.com/JFusco/cms-field-vocabulary/compare/v1.0.9...v1.0.10) (2026-09-26)

package/OPERATING.md

@@ -43,21 +43,21 @@ Never repair `catalog/`, `docs/platforms/`, or `dist/` directly.
 
 ## Vendor-documentation freshness
 
-The `Vendor documentation freshness` workflow runs an index scan daily at 13:17 UTC and runs an index scan plus the full official-source scan every Monday at 13:47 UTC. A manual dispatch can select either depth. It has read-only repository access plus permission to open or update issues.
+The `Vendor documentation freshness` workflow runs an index scan daily at 13:17 UTC and runs an index scan plus the full official-source scan every Monday at 13:47 UTC. A manual dispatch can select either depth. It has read-only repository access plus permission to open or update scanner-failure issues. A Codex project task reviews the weekly full scan and intervening daily observations every Monday at 15:00 UTC.
 
 The workflow:
 
 1. checks out the repository without persisted credentials;
 2. runs the release-index scan, adding the full official-source scan on the weekly schedule or a manual `full` dispatch;
 3. verifies that the scanner did not mutate the checkout;
-4. uploads logs and evidence for 30 days;
-5. opens or updates a source-specific attention issue when either scan reports drift or retrieval fails, and uses a scanner-integrity issue when the checkout changes.
+4. uploads logs, JSON observations, and a readable source-by-source summary for 30 days;
+5. opens an issue only if the scanner or summary fails unexpectedly or the checkout changes.
 
-It does not edit source, update `sources.lock.json`, open a vocabulary pull request, or publish.
+It does not edit source, update `sources.lock.json`, open a vocabulary pull request, or publish. The Codex task examines official evidence and proposes a focused draft pull request when a supported vocabulary change is clear. New major versions receive separate versioned profiles. It records a concrete finding and issue link with the draft. A reviewer accepts source observations and approves any merge; the task does not merge or publish.
 
-### Triage an attention issue
+### Triage a scan finding
 
-1. Open the linked workflow run and download the artifact.
+1. Open the latest completed workflow run and download the artifact.
 2. Inspect `index.log`, `full.log`, and `worktree-status.txt`.
 3. Identify each observation that is not `unchanged`.
 4. Open the registered official URL and review the exact locator in `sources/official-sources.json`.
@@ -91,7 +91,7 @@ pnpm run test:packed
 
 `pnpm sources:review -- --all` is reserved for an intentional complete-lock refresh. `--allow-manual` is an exception for an independently reviewed official source that automation cannot retrieve; the pull request must explain the failure and review method, and the reviewer must manually verify every required token. It cannot override a fetched observation that reports missing tokens or establish a new independent enumeration baseline.
 
-Close the attention issue only after the reviewed pull request is merged or the evidence establishes a transient false alarm. Link the issue to the reviewing pull request.
+Close a drift issue with a recorded no-impact finding, or after its reviewed pull request merges. The task must compare observation fingerprints with prior closed findings so a previously reviewed observation does not recreate the same issue. Link a change issue to its draft pull request. Leave a finding open when official evidence is missing, and state exactly what remains to verify.
 
 ## Review and release impact
 

package/README.md

@@ -38,6 +38,7 @@ Seven CMS identities are represented by sixteen versioned surface profiles:
 | Contentstack | `contentstack.cma.saas` | Management | SaaS snapshot observed 2026-09-01 |
 | Optimizely SaaS | `optimizely-saas.cms-api-v1` | Management | CMS SaaS API v1 |
 | Optimizely SaaS | `optimizely-saas.sdk-2` | SDK | `@optimizely/cms-sdk` 2.x contract snapshot |
+| Optimizely SaaS | `optimizely-saas.sdk-3` | SDK | `@optimizely/cms-sdk` 3.0.0 contract snapshot |
 | Optimizely PaaS | `optimizely-paas.cms12-model` | SDK/model | CMS 12 model and editor surface |
 | Optimizely PaaS | `optimizely-paas.cms12-property-data-type` | Storage | `EPiServer.dll` 12.0.3 enum |
 | Optimizely PaaS | `optimizely-paas.cms13-admin` | Authoring | CMS 13 admin labels |
@@ -153,9 +154,9 @@ The root export also provides catalog lookup and projection functions. Package s
 
 ## Freshness is reviewed, not auto-authored
 
-A scheduled, read-only workflow checks registered official release/version indexes daily and runs a full official-source scan weekly. It normalizes fetched documents, compares reviewed hashes, verifies required evidence tokens, and independently extracts complete token sets from structurally bounded official-source regions. The required-token check proves approved facts still have evidence; the independent observed-token hash detects native types added or removed without first adding them to the manifest. Claim text drift is classified separately when the observed enumeration is unchanged. Raw-only cosmetic churn remains visible in the uploaded report but does not create an attention issue.
+A scheduled, read-only workflow checks registered official release/version indexes daily and runs a full official-source scan weekly. It normalizes fetched documents, compares reviewed hashes, verifies required evidence tokens, and independently extracts complete token sets from structurally bounded official-source regions. The required-token check proves approved facts still have evidence; the independent observed-token hash detects native types added or removed without first adding them to the manifest. Claim text drift is classified separately when the observed enumeration is unchanged. Raw-only cosmetic churn remains visible in the uploaded report. The scan summary names affected profiles and shows added or removed release identities; ordinary drift no longer opens a generic GitHub issue.
 
-The scanner never edits vocabulary source, updates the reviewed lock, opens a vocabulary pull request, or publishes a package. A maintainer must inspect the official source at the recorded locator, update structured claims when warranted, run the explicit source-review command, and pass the full quality gate. See [CONTRIBUTING.md](./CONTRIBUTING.md) for evidence rules and [OPERATING.md](./OPERATING.md) for the freshness and release runbooks.
+The scanner never edits vocabulary source, updates the reviewed lock, opens a vocabulary pull request, or publishes a package. A scheduled Codex review can propose focused draft pull requests from scan evidence, but a maintainer must inspect the official source at the recorded locator and accept reviewed observations before merge. See [CONTRIBUTING.md](./CONTRIBUTING.md) for evidence rules and [OPERATING.md](./OPERATING.md) for the freshness and release runbooks.
 
 ## License and security