Repository navigation
Conversation
Bumps the devtools group with 1 update: [eslint](https://github.com/eslint/eslint). Updates `eslint` from 10.10.0 to 10.11.0 - [Release notes](https://github.com/eslint/eslint/releases) - [Commits](eslint/eslint@v10.10.0...v10.11.0) --- updated-dependencies: - dependency-name: eslint dependency-version: 10.11.0 dependency-type: direct:development update-type: version-update:semver-minor dependency-group: devtools ... Signed-off-by: dependabot[bot] <support@github.com>
Bumps [actions/checkout](https://github.com/actions/checkout) from 6 to 7. - [Release notes](https://github.com/actions/checkout/releases) - [Changelog](https://github.com/actions/checkout/blob/main/CHANGELOG.md) - [Commits](actions/checkout@v6...v7) --- updated-dependencies: - dependency-name: actions/checkout dependency-version: '7' dependency-type: direct:production update-type: version-update:semver-major ... Signed-off-by: dependabot[bot] <support@github.com>
The in-place page preview (--preview, --ocr, --serve) and XLIFF localization (--xliff, --dry-run, --max-segments) sat outside the skill and web surfaces, carried about 3,000 lines with almost no usage signal, and are not part of the core mission. Page preview may return in a future desktop app; the design lives in history before a18644b. Removed: src/preview/*, src/ocr.js, src/capped-fetch.js, src/cli/xliff.js, their seven unit test files and tests/fixtures/xliff/. args.js drops the six flags, the preview/serve/XLIFF validators, and the --preview-only comma-list variant compare: --jargon and --register now take exactly one value. run.js drops runPreviewJob/runOcrStage/runXliffMode; security.js drops the page-derived SSRF guard; resolveBatchOutputPath loses the XLIFF-only defaultSuffix option. Docs updated (CLI, FLAG-PARITY, ARCHITECTURE, EXIT-CODES, DESIGN, assets/demo/README). BREAKING CHANGE: six public CLI flags are gone; the next release is a major. Detection, scoring, prompts, the skill and the hosted surfaces are unchanged. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
With --ocr gone nothing passes images to a backend. Remove the `images` parameter on every invoke, `supportsImages`, `stageCliImages`, `selectOcrBackends`, the openai-http image rejection and doctor's `images=yes` marker. claude-cli now always runs with `--tools ""`, so AUTHENTICATION.md (+KR) no longer describes an OCR exception. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
`.patina.default.yaml` carried `stylometry.sentence_zoom.similarity_threshold`, `stylometry.skip.*` and `lexicon.skip.*`; no module in src/, api/ or bin/ reads them. The analyzer hardcodes the <=2 paragraph / <=2 sentence skip gate, and no surface implements a configurable zoom threshold, so core/stylometry.md §7 now states the 20% threshold as a rule constant. The parity test that pinned the yaml skip numbers to the analyzer goes with the keys. `.env.example` documented `PATINA_POLAR_LOCK_WAIT_MS`, which no code reads. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
…ompt section --jargon existed only on the CLI and never on the /patina skill, which broke the skill/CLI parity rule. Remove the flag, the explain/remove directives, the `jargon` option on buildPrompt/verifyRewrite, the hosted prompt-budget branch only a `jargon` request field could reach, and the rewrite-ab fixture field. The default behaviour stays: every rewrite prompt still tells the model to copy Latin-letter technical terms as-is (`buildTerminologyConstraint`, formerly the `--jargon keep` branch of buildTransformDirective). Prompt snapshots change only in the section header and bullet label, which drop the "(--jargon keep)" suffix. validateTransformRequest becomes validateRegisterRequest since register is the only option it guards. BREAKING CHANGE: the --jargon CLI flag is removed (exit 2, unknown option). Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
8.7.0 made the H-RHETORIC edit policy the default and kept the older similar-weight sentence behind PATINA_RHETORIC_POLICY=legacy. Nobody needs the rollback any more, so the switch goes: resolveRhetoricPolicy, the LEGACY_* prompt constants, and the rhetoricPolicy option on buildPrompt, verifyRewrite and buildWebRewritePrompt. The default prompt text is byte-identical (prompt snapshots unchanged). Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
SKILL.md 4.6 and core/stylometry.md §7 told the skill to group adjacent
same-length sentences inside a hot paragraph into `P{n}.S{m}-S{k}`
sub-flags. The CLI analyzer never produced them, so the two surfaces
disagreed. Remove the rule, its meta-block entries and worked-example
lines; 5b now prioritises hot paragraphs (`«P{n} SUSPECT»`). Section
numbers are left as they were (the doc already skipped §11).
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
feat!: remove --preview/--ocr/--serve and --xliff from the CLI
…tions/checkout-7 chore(deps): bump actions/checkout from 6 to 7
…ools-4ee7f9a52b chore(deps-dev): bump eslint from 10.10.0 to 10.11.0 in the devtools group
…ic switch (#952) chore: drop unread config keys and remove --jargon (Tier 1 cleanup)
The private pack repository has served an empty manifest since 2026-07-12, when its only pack was retired after Study 2, so the CLI command and the endpoint delivered nothing. `patina pack` now fails with the same "was removed" usage error as the retired `pattern` and `aside` commands (exit 2). Hosted Pro license validation and custom/ pattern and persona loading are unchanged.
Removes katfish-calibration (its regeneration was already retired), qa/pattern-overlap (no entry point or doc), qa/mdx-score (oh-my-docs loop closed), badge-json (duplicates patina-action's badge payload), and rewrite-ab with iterative-rewrite-baseline (parked for H-4a with no open issue; both are intact at 9108a63). None of them ship in the npm package.
* feat!: remove the kimi-cli backend Kimi Code takes the prose as a --prompt argv value that other local users can read from the process list, which is why the skill helper already refused it. `--backend kimi-cli` is now an unknown backend (exit 2) and `--model kimi-*` stays on openai-http. The Kimi/Moonshot HTTP presets and the hosted kimi provider stay; the research harness drops its kimi transport with the backend. * docs: drop Kimi CLI from the FAQ backend lists Review follow-up for #954: the English and Korean FAQ still offered a logged-in Kimi CLI as an API-key alternative. Also renames a sample backend error in batch.test.js.
Removes the browser analytics adapter, the same-origin /api/funnel collector, and every track() call site in the playground controller. Nothing in the repository read the Upstash counters and the campaign enum was still pinned to the 2026-09-07 launch. Server-side observability for pro-monitor, checkout UTM forwarding, and audit receipts are unchanged.
* release: prepare 9.1.0 — dead-feature sweep
- CHANGELOG 9.1.0 entry; minor by owner decision, with the removals
that strict semver would call major listed in the semver rationale
- version sync 9.0.0 -> 9.1.0 across package.json, package-lock.json,
SKILL.md, .patina.default.yaml, .claude-plugin/{plugin,marketplace}.json,
packages/patina-humanizer/package.json (version + patina-cli pin),
README{,_KR,_ZH,_JA}.md and the playground version badges
* release: name the --model kimi-* fallback in the 9.1.0 semver rationale
Review follow-up for #957.
Contributor
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
This branch was successfully deployed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Release 9.1.0: merge
devintomain. Approved by the owner on 2026-09-26. Policy:docs/WORKFLOW.mdIncluded, each already merged into
devwith green CI and one independent review (verdicts are on each PR):--preview/--ocr/--serveand--xliff/--dry-run/--max-segments--jargonandPATINA_RHETORIC_POLICY=legacy, plus unread config keyspatina pack,/api/packs)kimi-clibackendScope and risk
mainto production through the Vercel Git integration. Pushing the tagv9.1.0afterwards publishespatina-cliandpatina-humanizerto npm (Trusted Publishing) and creates the GitHub Release from the CHANGELOG. No GHCR image is published.Verification
devhead for this PR, 31dddb3):npm run release:check→ 0;npm run lint→ 0 (architecture 103 modules, 0 violations);npm test→ 0 (2018 tests, 2017 pass, 1 skipped);check:no-private-assets,dogfood, andbenchmark:report/benchmark:compare(timestamp-only drift) → 0;scripts/release-artifacts.mjs --dry-runverified both 9.1.0 tarballs.devpush CI green on every included merge commit.Release and rollback
patina-gb0s82zka(dpl_6fiQSeZVG5pTJFa4XV9d4sdB8KR2, 9.0.0).latestback withnpm dist-tag add patina-cli@9.0.0 latest(and the same forpatina-humanizer).