[AmitMY]

Adds a removable English learner extension to existing WordNet words and meanings. Missing examples appear through the normal example APIs; plain-language explanations and optional domain, register, usage, grammar and irregular-form notes are available through learner_data() and the web API. Original definitions and concept IDs stay unchanged.

The pinned top-5,000 frequency list reaches 21,043 existing senses. The release adds 14,048 example sentences to 6,424 concepts, plus 12,643 plain-language explanations and selective learner notes. 100 uncertain senses remain withheld, with specific reasons and recorded repair attempts.

Every published generated addition has a content-bound Astra approval. The complete initial audit checked 17,026 senses; subsequent corrections and independent re-reviews approved 2,685 repaired senses, including 83 previously withheld senses. The final gate validates 17,047 sense reviews. This is AI review, not human verification or a measured accuracy guarantee.

The release gate blocks missing, stale, negative, uncalibrated and self-authored approvals. Each reviewer passed 12 blind calibration cases. CI and Docker rebuild the XML from approved records to verify the packaged text. Original findings, rejected source, writing attempts and re-review decisions are preserved. Docker installation is opt-in with INSTALL_ENGLISH_LEARNER=true until the dictionary exposes AI attribution.

Validation: lint, mypy, package build, all 233 tests, and the full offline release gate pass. Installed against OMW English 1.4, checked API provenance, exported/reimported/removed the extension, and verified that original base rows and the user's database remained unchanged.

---

Note

Medium Risk
Large volume of AI-generated dictionary text enters user-facing APIs with automated—not human—review; regressions would affect lexical accuracy and attribution, though import remains additive and Docker install is opt-in.

Overview
Ships v1.9.0 with an additive English learner lexicon extension for frequent OMW English senses: AI-authored examples and scoped learner notes attach to existing words/senses/synsets without new concepts, exposed via learner_data() on core objects and matching web attributes.learner_data (original definitions unchanged).

Adds the full extensions/english-learner pipeline—inventory/generation, content-bound Astra review and calibration, build_extension.py, quarantine/correction helpers, committed rylo-en-learner.xml.gz plus manifest/audit artifacts, and check_release.py, which revalidates hashes, review coverage, withheld senses, and recompiles the packaged XML. CI and Docker now run that gate; Docker copies the extension and only runs wn add when INSTALL_ENGLISH_LEARNER=true (default off until AI attribution is shown in the product).

^{Reviewed by Cursor Bugbot for commit d0cb63a. Bugbot is set up for automated code reviews on this repo. Configure here.}

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit d0cb63a. Configure here.}

[cursor]

Corrections omit author-run binding

Medium Severity

Correction output sets requires_semantic_review but never writes author_run. The release gate only rejects a reviewer that matches that field, so the same Astra run can author a repair and then approve it. The self-authored-approval block described for this release therefore does not apply to generated corrections.

Additional Locations (1)

• extensions/english-learner/quality_gate.py#L235-L240

 

^{Reviewed by Cursor Bugbot for commit d0cb63a. Configure here.}