fix(research): require explicit invocation like the frontend skill

.claude-plugin/marketplace.json

@@ -9,7 +9,7 @@
       "name": "superloopy",
       "source": "./",
       "description": "Strict evidence loop harness: repo-local loop state, success criteria, append-only ledgers, evidence-gated completion, plus skills (loop, research, frontend, clone) and a read-only/worker crew. Works on Claude Code and Codex from one repo.",
-      "version": "0.11.0",
+      "version": "0.11.1",
       "license": "MIT",
       "keywords": [
         "claude-code",

.claude-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "superloopy",
-  "version": "0.11.0",
+  "version": "0.11.1",
   "description": "Lightweight loop harness with strict evidence gates — Claude Code edition.",
   "homepage": "https://github.com/beefiker/superloopy#readme",
   "repository": "https://github.com/beefiker/superloopy",

.codex-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "superloopy",
-  "version": "0.11.0",
+  "version": "0.11.1",
   "description": "Lightweight loop harness with strict evidence gates.",
   "homepage": "https://github.com/beefiker/superloopy#readme",
   "repository": "https://github.com/beefiker/superloopy",

README.es.md

@@ -46,7 +46,7 @@ Superloopy mantiene pequeña la capa de comandos. Las skills guardan el flujo es
 | --- | --- | --- |
 | `superloopy-loop` | Escribes `loopy <task>` o `loopy team <task>` para un loop completo; usa `loopywork`, `lpy` o `$lpy` para contexto solo de guidance. | Los loops completos producen un plan ligero, siguientes acciones guiadas, prueba respaldada por comandos, un quality gate y un evidence report final. Los alias de guidance no mutan estado. |
 | `superloopy-doctor` | Diagnosticas instalación, wrapper, plugin cache, hook/bootstrap, agentes, wiring de Codex/Claude Code o versiones stale. | Health report de solo lectura: evidencia de wrapper/cache/version, checks fallidos y el comando exacto de reparación que solo se ejecuta si lo apruebas. |
-| `superloopy-research` | Pides `loopy research`, deep research, exhaustive investigation o un informe con citas. | Ejes de investigación, expansion waves, un claim ledger, notas de verificación y un synthesis artifact citado. |
+| `superloopy-research` | Invocas explícitamente `$superloopy:superloopy-research` en Codex o `/superloopy:superloopy-research` en Claude Code, o inicias la investigación con un `loopy`/`루피` inicial (por ejemplo `loopy research`). Las peticiones simples de investigar, buscar o resumir no lo activan. | Ejes de investigación, expansion waves, un claim ledger, notas de verificación y un synthesis artifact citado. |
 | `superloopy-clone` | Pides `loopy clone`, clonación autorizada de un sitio, reconstrucción, migración o recuperación visual precisa de páginas. | Capturas de navegador, topología de página, design tokens, inventario de assets, notas de implementación, salida de build y evidencia de visual QA. |
 | `superloopy-frontend` | Solo para interfaces de aplicaciones en pantalla dentro del alcance —Web alojada en navegador (pública, autenticada, privada/interna, PWA instalada o extensión), Web interactiva de contenido ya desplegada (campañas, publicaciones o experiencias de landing con un recorrido de usuario), escritorio, móvil/tableta, clientes embebidos/híbridos, UI con renderizado personalizado, Qt o destinos mixtos—, invocas explícitamente `$superloopy:superloopy-frontend` en Codex o `/superloopy:superloopy-frontend` en Claude Code, o inicias ese trabajo con `loopy`/`루피` al principio. La mera terminología de UI, plataforma o framework no la activa; se excluyen TV, wearables, XR, automoción, UI de juegos, TUI, artefactos estáticos de medios/documentos y trabajo que no sea de UI. | Un contrato de UX compartido más rutas de plataforma/composición. La evidencia es proporcional a las afirmaciones modificadas y verifica por separado cada navegador, destino/shell nativo, renderizador y destino mixto. Las ejecuciones independientes conservan recibos por ejecución; los bucles activos los vinculan al goal y criterion. |
 | `humanize-korean` | Úsala cuando usuarios coreanos piden quitar el tono AI, corregir 번역투 o hacer que el texto coreano suene humano sin cambiar hechos. | Escribe `final.md`, `summary.md` y `audit.json`; en loops de Superloopy registra evidencia en `.superloopy/evidence/humanize-korean/`. |

README.ja.md

@@ -46,7 +46,7 @@ Superloopy はコマンド層を小さく保ちます。専門的な進め方は
 | --- | --- | --- |
 | `superloopy-loop` | 完全な loop は `loopy <task>` または `loopy team <task>` で始めます。`loopywork`、`lpy`、`$lpy` は guidance だけが必要なときに使います。 | 完全な loop は軽量な計画、次の行動、コマンドで検証した証拠、品質 gate、最終 evidence report を残します。Guidance alias は状態を変更しません。 |
 | `superloopy-doctor` | install、wrapper、plugin cache、hook/bootstrap、agent、Codex/Claude Code host wiring、stale version の問題を診断するとき。 | 読み取り専用の health report: wrapper/cache/version の証拠、失敗した check、承認後にだけ実行する正確な修復コマンド。 |
-| `superloopy-research` | `loopy research`、deep research、exhaustive investigation、または引用付きレポートを求めるとき。 | 調査軸、拡張 wave、claim ledger、検証メモ、引用付き synthesis artifact。 |
+| `superloopy-research` | Codex の `$superloopy:superloopy-research` または Claude Code の `/superloopy:superloopy-research` を明示的に呼び出すか、リサーチ作業を先頭の `loopy`/`루피`（例: `loopy research`）で始めたときのみ。単なる調査・検索・要約の依頼では起動しません。 | 調査軸、拡張 wave、claim ledger、検証メモ、引用付き synthesis artifact。 |
 | `superloopy-clone` | `loopy clone`、許可された Web サイトのクローン、再構築、移行、ピクセル単位の復元を求めるとき。 | ブラウザ取得、ページ構造、デザイントークン、アセット一覧、実装メモ、build 出力、visual QA 証拠。 |
 | `superloopy-frontend` | 対応範囲の画面ベースのアプリ UI（公開・認証済み・非公開/社内向け・インストール型 PWA/拡張機能を含むブラウザーホスト Web、ユーザージャーニーを備えたインタラクティブなデプロイ済みコンテンツ主導 Web（キャンペーン、出版、ランディング体験など）、デスクトップ、モバイル/タブレット、組み込み/ハイブリッドクライアント、カスタムレンダリング UI、Qt、混在ターゲット）の作業で Codex の `$superloopy:superloopy-frontend` または Claude Code の `/superloopy:superloopy-frontend` を明示的に呼び出すか、その作業を先頭の `loopy`/`루피` で始める場合だけ。単なる UI・プラットフォーム・フレームワーク用語では起動せず、TV、ウェアラブル、XR、自動車、ゲーム UI、TUI、静的なメディア/文書成果物、非 UI 作業は除外します。 | 1 つの共通 UX 契約にプラットフォーム/コンポジション経路を加えます。証拠は変更した主張に比例し、ブラウザー、ネイティブのターゲット/シェル、レンダラー、混在ターゲットの各所有者を独立して検証します。単独実行では実行単位の証跡を保持し、アクティブなループでは goal と criterion に関連付けます。 |
 | `humanize-korean` | 韓国語テキストの AI っぽさや翻訳調を抑え、事実を変えずに人が書いたように整えるとき。 | `final.md`、`summary.md`、`audit.json` を書き、Superloopy loop では `.superloopy/evidence/humanize-korean/` に証拠を残します。 |

README.ko.md

@@ -46,7 +46,7 @@ Superloopy는 명령을 작게 둡니다. 대신 스킬이 작업 방식을 잡
 | --- | --- | --- |
 | `superloopy-loop` | full loop는 `loopy <task>` 또는 `loopy team <task>`로 시작합니다. `loopywork`, `lpy`, `$lpy`는 가벼운 guidance만 넣을 때 씁니다. | full loop는 가벼운 계획, 다음 행동, 명령으로 검증한 근거, 품질 게이트, 최종 evidence report를 남깁니다. guidance alias는 상태를 바꾸지 않습니다. |
 | `superloopy-doctor` | install, wrapper, plugin cache, hook/bootstrap, agent, Codex/Claude Code host wiring, stale version 문제를 진단할 때. | 읽기 전용 health report: wrapper/cache/version 근거, 실패한 체크, 승인 후 실행할 정확한 복구 명령. |
-| `superloopy-research` | `loopy research`, deep research, exhaustive investigation, 출처 있는 리서치 보고서를 요청할 때. | 리서치 축, 확장 wave, claim ledger, 검증 메모, 출처가 붙은 synthesis artifact. |
+| `superloopy-research` | Codex의 `$superloopy:superloopy-research`나 Claude Code의 `/superloopy:superloopy-research`를 직접 호출하거나, 리서치 작업을 선행 `loopy`/`루피`(예: `loopy research`)로 시작할 때만. 단순한 조사·검색·요약 요청으로는 켜지지 않습니다. | 리서치 축, 확장 wave, claim ledger, 검증 메모, 출처가 붙은 synthesis artifact. |
 | `superloopy-clone` | `loopy clone`, 허가된 웹사이트 클론, 리빌드, 마이그레이션, 픽셀 기준 복구를 요청할 때. | 브라우저 캡처, 페이지 구조, 디자인 토큰, asset 목록, 구현 메모, build 출력, visual QA 근거. |
 | `superloopy-frontend` | 지원되는 화면 기반 애플리케이션 UI(공개·인증·비공개/내부용·설치형 PWA/확장을 포함한 브라우저 호스팅 웹, 사용자 여정이 있는 인터랙티브 배포형 콘텐츠 중심 웹(캠페인·출판·랜딩 경험 등), 데스크톱, 모바일/태블릿, 임베디드/하이브리드 클라이언트, 커스텀 렌더링 UI, Qt, 혼합 타깃) 작업에 Codex의 `$superloopy:superloopy-frontend`나 Claude Code의 `/superloopy:superloopy-frontend`를 직접 호출하거나, 해당 작업을 선행 `loopy`/`루피`로 시작할 때만. 단순한 UI·플랫폼·프레임워크 용어로는 켜지지 않으며 TV·웨어러블·XR·자동차·게임 UI·TUI·정적 미디어/문서 결과물·비 UI 작업은 제외됩니다. | 하나의 공통 UX 계약에 플랫폼/컴포지션 경로를 더합니다. 근거는 변경한 주장에 비례하며 브라우저, 네이티브 타깃/셸, 렌더러, 혼합 타깃의 각 소유자를 독립적으로 검증합니다. 독립 실행에서는 실행별 영수증을 보존하고 활성 루프에서는 goal과 criterion에 연결합니다. |
 | `humanize-korean` | 한국어 글의 AI 티를 줄이거나 번역투를 고치고, 사실은 바꾸지 않은 채 사람이 쓴 것처럼 다듬어야 할 때. | `final.md`, `summary.md`, `audit.json`을 쓰고, Superloopy loop 안에서는 `.superloopy/evidence/humanize-korean/` 아래에 근거를 남깁니다. |

README.md

@@ -46,7 +46,7 @@ Superloopy keeps the command layer small. Skills carry the specialist workflow:
 | --- | --- | --- |
 | `superloopy-loop` | You type `loopy <task>` or `loopy team <task>` for a full loop; use `loopywork`, `lpy`, or `$lpy` for guidance-only context. | Full loops produce a lightweight plan, guided next actions, command-backed proof, a quality gate, and a final evidence report. Guidance aliases do not mutate state. |
 | `superloopy-doctor` | You diagnose install, wrapper, plugin cache, hook/bootstrap, agent, Codex/Claude Code host wiring, or stale-version problems. | A read-only health report with wrapper/cache/version evidence, failing checks, and the exact repair command to run only if approved. |
-| `superloopy-research` | You ask for `loopy research`, deep research, exhaustive investigation, or a cited report. | Research axes, expansion waves, a claim ledger, verification notes, and a cited synthesis artifact. |
+| `superloopy-research` | You explicitly invoke Codex `$superloopy:superloopy-research` or Claude Code `/superloopy:superloopy-research`, or start the research task with a leading `loopy`/`루피` (such as `loopy research`). Plain research, investigate, or summarize requests do not activate it. | Research axes, expansion waves, a claim ledger, verification notes, and a cited synthesis artifact. |
 | `superloopy-clone` | You ask for `loopy clone`, authorized website cloning, rebuilding, migration, or pixel-focused page recovery. | Browser captures, page topology, design tokens, asset inventory, implementation notes, build output, and visual QA evidence. |
 | `superloopy-frontend` | You explicitly invoke Codex `$superloopy:superloopy-frontend` or Claude Code `/superloopy:superloopy-frontend` for supported screen-based application UI across browser-hosted Web (public, authenticated, private/internal, installed PWA, or extension), interactive deployed content-led Web (campaign, publication, or landing experiences with a user journey), desktop, mobile/tablet, embedded/hybrid clients, custom-rendered UI, Qt, or mixed targets, or start that work with a leading `loopy`/`루피`. Plain UI, platform, or framework terms do not activate it; TV, wearable, XR, automotive, game UI, TUI, static media/document artifacts, and non-UI work stay excluded. | One shared UX contract plus platform/composition routes. Evidence is proportional to changed claims and independently covers each browser, native target/shell, renderer, and mixed target. Standalone runs retain run-scoped receipts; active loops bind them to the goal and criterion. |
 | `humanize-korean` | Use when Korean users ask to remove AI tone, fix 번역투, or make Korean text sound human without changing facts. | Writes `final.md`, `summary.md`, and `audit.json`; in Superloopy loops it records evidence under `.superloopy/evidence/humanize-korean/`. |

README.zh-CN.md

@@ -46,7 +46,7 @@ Superloopy 保持命令层很小。具体工作方式由 skills 负责：什么
 | --- | --- | --- |
 | `superloopy-loop` | 用 `loopy <task>` 或 `loopy team <task>` 启动完整 loop；用 `loopywork`、`lpy`、`$lpy` 只注入 guidance。 | 完整 loop 会产出轻量计划、下一步指引、命令验证证明、质量 gate、最终 evidence report。Guidance alias 不会改状态。 |
 | `superloopy-doctor` | 诊断安装、wrapper、plugin cache、hook/bootstrap、agent、Codex/Claude Code host wiring 或版本过旧问题时。 | 只读 health report：wrapper/cache/version 证据、失败检查，以及只有批准后才运行的精确修复命令。 |
-| `superloopy-research` | 请求 `loopy research`、deep research、exhaustive investigation，或需要带引用的研究报告时。 | 研究轴、扩展 wave、claim ledger、验证笔记、带引用的 synthesis artifact。 |
+| `superloopy-research` | 仅当你在 Codex 中显式调用 `$superloopy:superloopy-research`、在 Claude Code 中调用 `/superloopy:superloopy-research`，或以行首 `loopy`/`루피`（如 `loopy research`）开始研究任务时。普通的调查、检索或总结请求不会激活它。 | 研究轴、扩展 wave、claim ledger、验证笔记、带引用的 synthesis artifact。 |
 | `superloopy-clone` | 请求 `loopy clone`、已授权的网站克隆、重建、迁移，或需要按像素恢复页面时。 | 浏览器截图、页面结构、设计 token、资产清单、实现笔记、build 输出、visual QA 证据。 |
 | `superloopy-frontend` | 仅在处理受支持的基于屏幕的应用 UI（浏览器托管 Web，包括公开、需认证、私有/内部、已安装 PWA 或扩展；具有用户旅程的已部署交互式内容型 Web，例如营销活动、出版物或着陆页体验；以及桌面、移动设备/平板、嵌入式/混合客户端、自定义渲染 UI、Qt 或混合目标）时，在 Codex 中显式调用 `$superloopy:superloopy-frontend`，或在 Claude Code 中调用 `/superloopy:superloopy-frontend`，也可用开头的 `loopy`/`루피` 启动该工作。仅出现 UI、平台或框架词汇不会激活它；TV、可穿戴设备、XR、汽车、游戏 UI、TUI、静态媒体/文档产物和非 UI 工作仍被排除。 | 采用一份共享 UX 契约，并叠加平台/界面构成路径。证据与变更声明成比例，并分别验证浏览器、原生目标/外壳、渲染器和每个混合目标。独立运行保留按次划分的证据；活动循环则把证据绑定到 goal 和 criterion。 |
 | `humanize-korean` | 需要去掉韩文内容里的 AI 腔、修正翻译腔，或在不改事实的前提下让韩文更像真人写作时。 | 写入 `final.md`、`summary.md`、`audit.json`；在 Superloopy loop 中把证据记录到 `.superloopy/evidence/humanize-korean/`。 |

docs/superloopy-file-audit.md

@@ -333,8 +333,8 @@ Superloopy is its own lightweight loop harness: one small CLI, repo-local `.supe
 | `skills/superloopy-frontend/scripts/visual-diff.mjs` | Dependency-free PNG visual-diff for the frontend visual-QA gate: similarityScore, alpha intactness, and ranked grid hotspots as review evidence. | Superloopy-native; uses only node:zlib/node:fs, no external image library or vendored code. |
 | `skills/superloopy-loop/SKILL.md` | Skill instructions for Superloopy guide, proof, capture, evidence, check, finish, review, checkpoint, and doctor flow. | Uses Superloopy-native workflow terms. |
 | `skills/superloopy-loop/agents/openai.yaml` | Minimal agent metadata for Superloopy discovery. | Search terms are Superloopy-native. |
-| `skills/superloopy-research/SKILL.md` | Skill instructions for deep research with EXPAND waves, claim verification, cited synthesis, and Superloopy evidence artifacts. | Superloopy-native research workflow; workers return text and parent-owned artifacts remain authoritative. |
-| `skills/superloopy-research/agents/openai.yaml` | Minimal agent metadata for discovering the Superloopy research skill. | Superloopy-native skill metadata only. |
+| `skills/superloopy-research/SKILL.md` | Explicit-invocation-gated skill instructions for deep research with EXPAND waves, claim verification, cited synthesis, and Superloopy evidence artifacts. | Superloopy-native research workflow; workers return text and parent-owned artifacts remain authoritative. |
+| `skills/superloopy-research/agents/openai.yaml` | Minimal explicit-only agent metadata for discovering the Superloopy research skill without implicit activation. | Superloopy-native skill metadata only. |
 | `skills/superloopy-slides/LICENSE` | Upstream MIT license for the adapted frontend-slides skill content. | License text only; records third-party provenance and no executable plugin logic. |
 | `skills/superloopy-slides/SKILL.md` | Skill instructions for zero-dependency HTML slide decks: style discovery, fixed 16:9 stage, PPT conversion, PDF export, and a visual-QA evidence gate. | Superloopy-native evidence gate on `.superloopy/evidence/slides/`; adapted from the MIT frontend-slides skill with attribution. |
 | `skills/superloopy-slides/STYLE_PRESETS.md` | Twelve curated slide style presets with palettes, fonts, and signature elements. | Design reference content for deck generation only; no executable plugin logic. |