feat: close im-not-ai competitive gaps — plugin marketplace, multi-agent --strict, KO scholarship (5.1.0)#497
Merged
Merged
Conversation
…sion-sync
Add .claude-plugin/{plugin,marketplace}.json so `/plugin marketplace add
devswha/patina` then `/plugin install patina@patina` works; the repo-root
SKILL.md auto-loads as the /patina skill (no skills/ dir, no skills manifest
field), so pattern/core/lexicon data dirs are untouched. Add uninstall.sh for
script-install parity (symlink/clone removal with safety guards). Wire both
manifests into scripts/check-release-metadata.mjs so their version stays synced
with package.json. Document the plugin path as INSTALLATION.md "Path 0".
Add three read-only subagents auto-discovered by the plugin: patina-detector (stylometry + AI-lexicon + pattern findings), patina-fidelity-auditor (MPS/fidelity audit), patina-naturalness-reviewer (residual-tell + over-edit grade). Add an opt-in --strict multi-pass mode to SKILL.md: P1 detect -> P2 rewrite -> P3 fidelity/MPS audit -> P4 naturalness re-scan -> P5 accept/retry/rollback. In a plugin environment --strict delegates the read-only analysis passes (P1/P3/P4) to the subagents via the Task tool (added to allowed-tools); otherwise it runs the same passes inline in a single agent. Identical floors/gate logic in both modes. docs/agents.md documents the flow and auto-delegation. No asserted numeric line or version (5.0.0) changed.
…version Add docs/research/ko-translationese-scholarship.md anchoring patina's Korean detectors to verifiable sources (Gellerstam 1986, Baker 1993, Toury 1995, Toral 2019; KO tradition: 이근희 2005/2008, 김순영 2012) with an honest scope note (concept/terminology grounding, not per-rule derivation; signals stay advisory and density-gated). Link it from TRANSLATIONESE-KO.md. Add the closest open-source skill (epoko77-ai/im-not-ai) to COMPARISON.md with a balanced, factual comparison. Fix stale ROADMAP version metadata (4.0.0 -> 5.0.0).
…re links Add the plugin-marketplace install path to all locales and fix stale version badges (KR/ZH/JA 4.0.1 -> 5.0.0). Rework README_KR quick-start for non-technical users: orientation note, natural-language usage (method A) vs slash command (method B), and a "say it in plain Korean -> patina option" mapping table (incl. --strict). Add scholarship + subagents/strict-flow doc links (EN + KR).
Minor bump for the additive plugin-marketplace distribution, the opt-in multi-agent --strict skill mode, and the Korean translationese scholarship docs. No removals/breaking; deterministic engine unchanged; all new behavior opt-in. Sync all version-bearing surfaces (package.json, patina-humanizer alias, SKILL.md, .patina.default.yaml, README + locale badges, .claude-plugin manifests) and add the CHANGELOG 5.1.0 entry. `release:check` green.
Contributor
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
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
Closes the competitive gaps identified against the closest open-source skill (
epoko77-ai/im-not-ai): distribution reach, multi-agent architecture, Korean academic authority, and non-technical onboarding. Bumps to 5.1.0 (minor — additive, opt-in, no removals; deterministic engine unchanged).Commits
/plugin marketplace add devswha/patina→/plugin install patina@patina; the repo-rootSKILL.mdauto-loads as the/patinaskill, data dirs untouched),uninstall.sh, and manifest version-sync wired intorelease:check.patina-detector/patina-fidelity-auditor/patina-naturalness-reviewer) auto-discovered fromagents/, plus an opt-in--strictmulti-pass mode (detect → rewrite → fidelity/MPS audit → naturalness re-scan → accept/retry/rollback). In a plugin env--strictdelegates the read-only passes to the subagents viaTask; otherwise it runs them inline in a single agent with identical floors/gate logic.docs/research/ko-translationese-scholarship.mdanchoring KO detectors to verifiable sources (Gellerstam 1986, Baker 1993, Toury 1995, Toral 2019; 이근희 2005/2008, 김순영 2012) with an honest advisory-only scope note;im-not-aiadded todocs/COMPARISON.md; ROADMAP version fix.Verification
npm test— 782 pass / 0 failnpm run lint— syntax OK, cspell 0 issuesnpm run release:check— OK for 5.1.0npm run check:no-private-assets— OKAll new behavior is opt-in (default-off), so existing runs are byte-identical without the new flags.
--strictis skill-only (not a CLI flag). The pre-publishbenchmark:report/dogfoodgates (LLM-backed) were not run here.