cliff-stack
v1.1.0Stack complet : 24 skills, 14 agents, 6 hooks (guards securite, son, timer, phpcs) et assets
By Cliff0 GitHub starsUpdated 12 hours ago
Directory evidence
- Runtimes
- Claude Code
- Parsed components
- 22 skill or MCP entries
- Source updated
- Sep 23, 2026
- Manifest status
- Canonical path parsed
The directory validates manifest shape and source location. It does not execute the plugin or provide a security endorsement. Review the indexing methodology →
Install cliff-stack for Claude Code
claude plugin marketplace add IchenDEV/agent-plugin-mkt
claude plugin marketplace update agent-plugin-marketplace
claude plugin install cliff-stack@agent-plugin-marketplacePaste and run these commands in a terminal with Claude Code. They add and refresh the PluginsMP catalog, then install this plugin.
The installer fetches third-party code from the source repository shown on this page. This directory validates manifest structure and source location, but does not perform a security audit; review the manifest, components, and source before installing.
Get the source manually
git clone https://github.com/Cliff-FK/.claudeClone the source repository, then follow its setup instructions to add the plugin to a compatible client. The plugin root is skills/cliff-stack/.
Plugin files
├── .claude-plugin/plugin.json├── skills/agent-builder/SKILL.md├── skills/asana-triage/SKILL.md├── skills/astro-native/SKILL.md├── skills/cancellation-flow-designer/SKILL.md├── skills/copywriting-landing/SKILL.md├── skills/design-auditor/SKILL.md├── skills/design-motion-principles/SKILL.md├── skills/email-lifecycle/SKILL.md├── skills/fable-mode/SKILL.md├── skills/freemius/SKILL.md├── skills/frontend-design/SKILL.md├── skills/impeccable/SKILL.md├── skills/marketing-psychology/SKILL.md├── skills/minify-assets/SKILL.md├── skills/paid-ads/SKILL.md├── skills/pricing-strategist/SKILL.md├── skills/seo-launch/SKILL.md├── skills/skill-builder/SKILL.md├── skills/wp-incident-response/SKILL.md├── skills/wp-native/SKILL.md├── skills/wp-plugin-check/SKILL.md└── skills/wp-save-ui-test/SKILL.md
Included Skills22
Use when the user asks to create, build, design or improve a Claude Code subagent ("crée/fais/conçois un agent sur X", "un agent spécialisé Y", "un sous-agent qui Z", "améliore/mets à jour cet agent"). Combines Anthropic's official subagent doctrine (single-responsibility, description = when-to-delegate + triggers + "use PROACTIVELY", least-privilege tools, model-by-cognitive-load, focused structured system prompt) with proven house conventions (runtime discovery / nothing hardcoded, breaks_if_touched alarm list, cross-zone links, anti-"resolved"-by-proxy, a binding finding-contract for finding-producing agents, semantic density not terseness) AND a mandatory adversarial challenge — a REAL critic subagent spawned to INVALIDATE it (mis-triggering, overlap with existing agents, wrong tool/model scope) plus an eval baseline — before delivery. Produces an agent under ~/.claude/agents/ (or .claude/agents/ for a project agent).
Triage industrialisé des tickets Asana de l'utilisateur. Détecte ses tickets ASSIGNÉS non terminés (tous les workspaces, identité et workspaces découverts dynamiquement), les comprend (texte + commentaires + pièces jointes images/PDF analysées visuellement), identifie par preuve le dossier projet local sous la racine web locale, produit un rapport + un début de correctif par ticket — PUIS, sur validation explicite « go » de l'utilisateur, lance les corrections en fan-out (1 agent par correction, max 5 en parallèle) avec validation adversariale et test Playwright réel si pas sûr à 100 %. À utiliser quand l'utilisateur demande de « voir / lister / traiter / triager mes tickets Asana », « qu'est-ce qui m'est assigné », « analyse mes tickets », ou veut démarrer les correctifs de ses tickets. Le pipeline lit en Asana et peut poster une relance de commentaire sur un ticket incompris ; il ne corrige le code QUE sur go explicite.
Use when implementing or modifying ANY Astro feature on an Astro project — .astro components, islands, client/server directives, Content Collections, Actions, view transitions, SSR adapters, i18n, middleware. Enforces Astro best practices: zero-JS-by-default, static-first then minimal hydration, Content Collections over raw fs, APIs verified via Context7, perf/LCP as a hard guardrail. Triggers on "ajoute/code une feature/un composant/une page/une island/une collection Astro".
Designs the subscription CANCELLATION / retention flow for a freemium WordPress plugin or SaaS — the 'reason → matched offer → confirmation' sequence (survey the cancellation reason, present ONE offer matched to that reason: pause / downgrade / support / last-resort one-time coupon), with save-rate analytics and HARD legal/dark-pattern guardrails. Decides the flow STRATEGY and copy; delegates the licensing/coupon/pause wiring to freemius and the dark-pattern/a11y audit to design-auditor. Use when asked to design a cancellation flow, reduce voluntary churn at the cancel moment, build a retention/offboarding/win-on-cancel flow, add a cancellation survey with offers, decide what to offer when someone cancels, or 'how do I keep customers from cancelling'. NOT failed-payment/dunning recovery (that's involuntary churn → freemius), NOT win-back emails after they've left (use email-lifecycle), NOT the price/tier design (use pricing-strategist).
Rédige et structure le MESSAGE d'une page de vente qui convertit pour un produit logiciel (plugin WordPress, thème, SaaS prosumer/B2B) : headline above-the-fold, ordre des sections (hero→problème→solution→preuve→pricing→CTA), copy des fonctionnalités (FAB), microcopy & CTA, traitement des objections. Choisit le bon framework selon la conscience de l'audience (PAS / PASTOR / AIDA). À utiliser quand on demande d'écrire/structurer/améliorer une landing, page de vente, page produit, hero, headline, accroche, CTA, argumentaire, copy pour vendre un plugin/SaaS/thème. Distinct de design-auditor (qui AUDITE le rendu), de marketing-psychology (les biais sous-jacents) et de pricing-strategist (le niveau de prix). NOT la microcopy/CTA d'upgrade affichée DANS l'éditeur Gutenberg → in-product-upgrade-prompts ; NOT les emails de conversion → email-lifecycle.
Audit UI/design across Figma and code (HTML/CSS/React/Vue/Tailwind): accessibility (WCAG/a11y/contrast/color-blindness), responsive, design tokens, forms, microcopy, Nielsen heuristics, AND dark patterns / ethical design / GDPR compliance. Outputs before/after diffs and dev handoff. Triggers on: review my UI, audit my design, is this accessible, color contrast, WCAG, a11y, Figma audit, dark patterns, is this manipulative, is this GDPR compliant, review my checkout, heuristic review, wireframe to spec. NOT the iterative building/improvement/polish of an interface (use impeccable) — this skill produces the formal audit report, it does not craft the UI. NOT the motion design authority (use design-motion-principles) — it can flag motion issues during an audit but defers to that skill for motion doctrine.
Motion and interaction design expert based on Emil Kowalski, Jakub Krehel, and Jhey Tompkins' techniques. Two modes — build interactive components with purposeful motion, or audit existing animations to catch AI-slop motion patterns (audit emits a branded HTML report with looping demos). Use when creating, adding, animating, or reviewing UI motion: transitions, hover states, micro-interactions, enter/exit animations, or any motion design work in React, Framer Motion, CSS, or HTML. Provides per-designer perspectives with context-aware weighting. The reference motion authority of this setup — other design skills defer to it for motion doctrine. NOT the general design/a11y/ethics audit (use design-auditor). NOT building or improving a complete UI (use impeccable) — this skill owns the motion layer only.
Conçoit les SÉQUENCES email lifecycle pour un SaaS / plugin WordPress freemium (éditeur solo/petite équipe) : welcome/onboarding vers le aha moment, activation, upgrade free→paid, trial-to-paid, nurturing, win-back/réactivation. Triggers comportementaux, segmentation par lifecycle stage, métriques fiables post-Apple-MPP, délivrabilité Gmail/Yahoo, conformité GDPR (marketing vs transactionnel), coordination avec le dunning du billing. À utiliser quand on demande une séquence/automation email, onboarding, nurturing, conversion d'essai, relance, réactivation, drip campaign, newsletter produit, ou une stratégie email pour convertir/retenir des utilisateurs d'un plugin/SaaS. Le dunning (relance de paiement échoué) reste géré côté Freemius — voir [[freemius]]. NOT l'upsell free→pro affiché DANS l'éditeur Gutenberg → in-product-upgrade-prompts ; ce skill ne couvre la conversion free→paid que PAR EMAIL.
Calibrage comportemental du modèle Claude Fable 5 / Mythos 5 pour éliminer ses dérives documentées et brancher les méthodologies, skills et agents maison, sans brider sa capacité. Use when l'utilisateur invoque /fable-mode, bascule sur Fable 5 / Mythos 5 via /model pour travailler, ou attaque SOUS Fable une tâche longue, agentique, multi-zone ou à fort enjeu (refactor, debug intermittent, migration, audit, run autonome). Déclencheurs : "/fable-mode", "mode fable", "calibre fable", "je passe sur fable", "je travaille sous fable", "run autonome sous fable", "tâche longue sous fable". NE PAS déclencher sur une simple question AU SUJET de Fable (comparaison de modèles, prix, capacités), ni sous Opus/Sonnet/Haiku (dérives différentes), ni sur une tâche triviale.
Use for ANYTHING Freemius (freemius.com) — the monetization platform for WordPress plugins/themes, SaaS and apps. Covers BOTH code (the @freemius/sdk JS/TS SDK, freemius-php-sdk, freemius-node-sdk, WordPress SDK, Checkout JS, webhooks, REST API, entitlements, customer portal) AND product/strategy (plans & pricing, subscriptions, trials, licensing, multi-currency, VAT/US sales tax, dunning, coupons, affiliate, refunds, deployments/releases). Enforces: real SDK surface (never invented field names), Context7 as source of truth before coding, security guardrails (HMAC webhook, server-only secrets, no IDOR). Triggers on "Freemius", "@freemius/sdk", licence/abonnement/webhook/checkout/portal Freemius, entitlements, déploiement/release Freemius. NOT le "vendre/monétiser mon plugin" générique sans Freemius nommé — modèle de prix/paliers → pricing-strategist, copy de page de vente → copywriting-landing, acquisition → seo-launch/paid-ads.
Guidance for distinctive, intentional visual design when building new UI or reshaping an existing one. Helps with aesthetic direction, typography, and making choices that don't read as templated defaults. Use when defining the art direction or aesthetic identity of a NEW interface or page, pairing display/body typefaces and setting a type scale, choosing a palette/layout with a point of view, or when a design risks reading as a generic template ('make it distinctive', 'give it personality', 'it looks AI-generated'). NOT the formal accessibility/WCAG/GDPR/dark-pattern audit (use design-auditor). NOT the iterative improvement/polish workflow on an existing interface (use impeccable). NOT motion/animation doctrine (use design-motion-principles).
Use when the user wants to design, redesign, shape, critique, audit, polish, clarify, distill, harden, optimize, adapt, animate, colorize, extract, or otherwise improve a frontend interface. Covers websites, landing pages, dashboards, product UI, app shells, components, forms, settings, onboarding, and empty states. Handles UX review, visual hierarchy, information architecture, cognitive load, accessibility, performance, responsive behavior, theming, anti-patterns, typography, fonts, spacing, layout, alignment, color, motion, micro-interactions, UX copy, error states, edge cases, i18n, and reusable design systems or tokens. Also use for bland designs that need to become bolder or more delightful, loud designs that should become quieter, live browser iteration on UI elements, or ambitious visual effects that should feel technically extraordinary. Not for backend-only or non-UI tasks. Not for formal accessibility/GDPR/dark-pattern audit reports — use design-auditor.
When the user wants to apply psychological principles, mental models, or behavioral science to marketing. Also use when the user mentions 'psychology,' 'mental models,' 'cognitive bias,' 'persuasion,' 'behavioral science,' 'why people buy,' 'decision-making,' or 'consumer behavior.' This skill provides 70+ mental models organized for marketing application. NOT pricing-model/tier DESIGN or WTP/Van Westendorp analysis (use pricing-strategist) — it only supplies the behavioral framing layer (charm/decoy/anchoring) once tiers exist. NOT writing/restructuring landing or email copy (use copywriting-landing/email-lifecycle) — it diagnoses which biases are missing, it does not produce the page copy.
Minifie les JS/CSS d'un projet (n'importe lequel) en fichiers .min via esbuild (npx, zéro dépendance projet), avec validation 3 niveaux (syntaxe, AST, équivalence fonctionnelle) et rapport de gains. À utiliser quand l'utilisateur demande de minifier/compacter/optimiser des assets JS ou CSS, de générer des .min, ou de préparer un dist. Ne branche JAMAIS les .min dans le code (enqueues/imports/HTML) sans validation explicite.
Produit le PLAN d'acquisition payante pour vendre un produit logiciel (plugin WordPress, thème, SaaS prosumer/B2B) à petit budget : choix de canal (Google Search/Bing d'abord), structure de compte/campagnes, angles créa (RSA, démo), plan budget/enchères (Smart Bidding & seuils), framework de mesure ROAS/CAC/LTV et attribution cookieless (Consent Mode v2). À utiliser quand on demande une stratégie/un plan ads, Google Ads, Meta/Facebook ads, campagne payante, SEA, PPC, retargeting, acquisition payante, ou 'dois-je faire de la pub pour mon plugin/SaaS'. Produit un PLAN (structure, créas, budget, mesure) — il ne pilote PAS une campagne live. Dit honnêtement quand le paid est un piège. Distinct de seo-launch (organique) et copywriting-landing (la page d'atterrissage).
Use when designing or revisiting product pricing — selecting a pricing model (subscription seat-based, usage-based, value-based, freemium, or hybrid), running Van Westendorp Price Sensitivity Meter analysis on WTP survey data, or designing Good/Better/Best packaging tiers. Recommends a model and a price range with trade-offs, never a single number. For Commercial leads, Product Marketing, and CMOs at the pricing-design moment — not deal-by-deal discounting, not brand positioning.
Stratégie SEO & contenu pour acquérir du trafic organique DURABLE pour un produit logiciel (plugin WordPress, thème, SaaS prosumer/B2B) : recherche de mots-clés intent-first, architecture topic-clusters + money pages, SEO technique on-page (Core Web Vitals, données structurées), E-E-A-T pour éditeur inconnu, et AEO (être cité par les moteurs IA / AI Overviews). À utiliser quand on demande une stratégie SEO, du contenu/blog pour ranker, des mots-clés, une page 'alternative à X' / comparatif, du référencement, de l'acquisition organique, de l'optimisation pour Google ou les réponses IA, un plan de lancement SEO. Distinct de copywriting-landing (le message de conversion) et de paid-ads (l'acquisition payante). NOT le classement dans le moteur de recherche interne du dépôt WordPress.org (readme.txt/tags/active installs/discoverability sur le repo) → wporg-readme-optimizer ; seo-launch = Google/web/AEO uniquement.
Use when the user asks to create, build, design or improve a Claude Code skill on a given subject ("crée/fais/conçois un skill sur X", "un skill spécialisé Y", "à la pointe sur Z", "améliore ce skill"). Combines Anthropic's official skill-authoring doctrine (eval-driven authoring with baseline, progressive disclosure, ≤500-line SKILL.md, third-person description, bundled scripts, multi-model testing, security audit) with a proven house research method (adversarial verification of "doesn't exist" claims, ground-truth over web over memory, parallel fan-out, encode the gotchas not the docs). Produces a global skill under ~/.claude/skills/ by default. NOT pour un sous-agent délégable à contexte isolé (avec ses propres tools restreints) → agent-builder ; ce skill ne fabrique qu'une procédure/connaissance chargée dans le contexte principal.
Use when a WordPress site is hacked, compromised, infected or defaced ("site WordPress piraté/hacké/infecté", "malware WordPress", "site nettoyé qui se réinfecte"), shows SEO spam or redirects to spam for Google visitors ("redirige vers du spam", "spam SEO", "cloaking"), has unknown or fake admin accounts ("des admins inconnus", "80 comptes administrateurs"), backdoors/webshells, a host or scanner flag, or needs a post-breach desinfection without SSH (panel + FTP, dropped PHP script) or offline (SQL dump + files), AND when WordPress or plugin updates are stuck/blocked/impossible ("les mises à jour ne se font plus", "auto-update désactivé"). Read-only investigation first, then guided containment, cleanup, secret rotation, hardening and verification. NOT for building a WordPress feature (wp-native) nor the sellable-plugin compliance gate (wp-plugin-check).
Use when implementing or modifying ANY WordPress feature on a WordPress project — Gutenberg blocks, block editor extensions (InspectorControls, HOC/filters), render_block/PHP hooks, theme.json, ACF blocks, REST, admin UI. Enforces 2026 WP-native best practices: core blocks + block supports first, native filters/HOCs over custom, @wordpress/* React components (never raw React), APIs verified via Context7, official coding standards. Triggers on requests like "ajoute/code une feature/un bloc/un réglage/un panneau/une option WP|Gutenberg|ACF". Inclut l'IMPLÉMENTATION éditeur d'un upsell freemium une fois la stratégie posée (Placeholder Pro dans edit(), registerBlockVariation scope:['inserter'], Pro-badge sur contrôle désactivé) — la STRATÉGIE de placement/timing reste à in-product-upgrade-prompts, le gating de licence à freemius. NOT la procédure de vrai save humain éditeur (clic Enregistrer réel pour déclencher cache/rest_after_insert/preSavePost) → wp-save-ui-test.
Gate "plugin vendable" — lance l'outil officiel WordPress Plugin Check (wp plugin check <slug>) sur un plugin WP et synthétise les violations (sécurité, i18n, perf, a11y, conformité repo WordPress.org). À utiliser AVANT de livrer/commiter un plugin destiné à être distribué ou vendu, ou quand on demande "vérifie/audite la conformité du plugin", "est-ce que le plugin passe Plugin Check", "prépare le plugin pour le repo". NE PAS utiliser en per-édition (l'outil scanne tout le plugin, ~10 s) — c'est un gate ponctuel, le lint per-fichier reste PHPCS. Ce skill SCANNE et synthétise les violations ; l'implémentation des correctifs (esc_*, sanitize, nonce, textdomain) revient à wp-native.
Procédure pour déclencher un VRAI save humain dans l'éditeur WordPress/Gutenberg via Playwright (clic réel sur Enregistrer, jamais programmatique), afin de tester un comportement qui ne se déclenche qu'au save UI (cache plugin, hooks rest_after_insert, filtres preSavePost, validation REST). À utiliser quand : on debug/teste un plugin ou thème WordPress dont le comportement dépend d'une sauvegarde réelle de post depuis l'admin, et qu'un save programmatique (wp post update / savePost) ne reproduit pas le cas. Ne s'applique qu'aux projets WordPress.
Plugin manifests1
{
"$schema": "https://json.schemastore.org/claude-code-plugin-manifest.json",
"name": "cliff-stack",
"displayName": "Cliff Stack",
"version": "1.1.0",
"description": "Stack complet : 24 skills, 14 agents, 6 hooks (guards securite, son, timer, phpcs) et assets",
"author": {
"name": "Cliff"
}
}For maintainers
If you maintain this plugin, link to this source-backed listing from your README so users can review its manifest and indexed components.
[cliff-stack on Agent Plugins Marketplace](https://pluginsmp.com/plugins/cliff-stack)