Recently added
Plugins grouped by the day they were added to the directory, newest first.
Sep 15, 2026
- dsh-skills-anywhere6stars
Discover and load Agent Skills from configured local directories, plugin marketplaces and Git sources through a local stdio MCP server.
Claude CodeAgent Plugins1 MCP server - deepr-research10stars
Bounded research and persistent expert discovery through an Agent Skill and a local read-only MCP capability profile.
Agent Plugins1 skill1 MCP server - golf2stars
Golf Intelligence, by Stracka. The highest-quality proprietary golf course dataset for developers building a golf app. Continuously updated mapped course data — not a scrape. Search is free. Scorecards, GPS, and 3D greens via API.
Claude CodeAgent Plugins1 skill1 MCP server - housecarl32stars
Comprehensive data-layer access to your Skyrim SE load order — read any record, author patches into a new plugin, look mods up on Nexus, and look up record schemas, Papyrus signatures, performance reviews, and distributor grammars — in plain English.
Claude Code1 MCP server - shunt2.3kstars
Shunts I/O-heavy work to AiKA modes to save tokens. Hooks block large file reads and redirect to bulk-reader/code-writer scripts backed by the Portal CLI.
Claude Code2 skills - portal2.3kstars
Set up, diagnose, search, and operate Spotify Portal through focused workflows and the Portal CLI.
CodexClaude Code6 skills - blunt0stars
Injects blunt, outcome-first communication guidelines at session start.
Claude Code - dgnengine0stars
엔솔 지식 그래프(Neo4j) 읽기 전용 조회 — ensol-kg 스킬 + MCP 접속 + /kg-doctor 연결 점검
Claude Code1 skill - aaron-marketing0stars
120 marketing skills and 8 commands as an AI marketing staff — one plugin, portable skills on 70+ hosts, or an 8-bot team — and as a four-layer operating system: Narrative strategy; SEO/GEO, Organic Social, Email, Paid Ads, and Influencer channels; Product Launch orchestration; and a Protocol layer with seven append-only truth registries plus non-canonical working memory. Eight typed advisory frameworks (TALE, CORE-EEAT, CITE, STAR, ECHO, SEND, ROAS, RAMP) use machine-checkable auditor artifacts and immutable standalone runtime contracts. Keyless-first and zero-dependency.
Claude Code - yandex-wordstat2stars
Анализ поискового спроса (Wordstat). ⚠️ Доступ к scope требует ручной заявки в поддержку Яндекса.
Claude Code1 skill - yandex-webmaster2stars
Управление сайтами в Вебмастере: индексация, поисковые запросы, сайтмапы, переобход, ссылки.
Claude Code1 skill - yandex-tracker2stars
Трекер задач: задачи, очереди, комментарии, статусы, отчёты.
Claude Code1 skill - yandex-telemost2stars
Видеовстречи Яндекс.Телемост: создание встреч, расписание, ссылки.
Claude Code1 skill - yandex-metrika2stars
Аналитика Яндекс.Метрики: трафик, конверсии, цели, UTM, источники, e-commerce, расходы Директа. Cache-first, 30-строчный лимит stdout.
Claude Code1 skill - yandex-messenger2stars
Яндекс.Мессенджер: чаты, сообщения, поиск (бот-API).
Claude Code1 skill - yandex-mail2stars
Почта Яндекс 360: чтение, отправка, поиск через IMAP/SMTP с OAuth-токеном.
Claude Code1 skill - yandex-forms2stars
Конструктор форм: создание форм, чтение ответов, экспорт результатов.
Claude Code1 skill - yandex-disk2stars
Облако Яндекс.Диск через REST API: файлы, папки, публикация ссылок. Также fallback на WebDAV.
Claude Code1 skill - yandex-direct2stars
Контекстная реклама Яндекс.Директ: кампании, объявления, отчёты, ставки, ОРД-документы.
Claude Code1 skill - yandex-calendar2stars
Календарь Яндекс 360 через CalDAV: события, расписание, напоминания.
Claude Code1 skill - yandex-bot-platform2stars
Ботоплатформа Яндекс.Диалогов: навыки Алисы, голосовые скиллы, аналитика.
Claude Code1 skill - yandex-auth2stars
Unified OAuth flow for all Yandex APIs (single token covering all scopes the user enabled in their Yandex OAuth app)
Claude Code1 skill - yandex-audience2stars
Аудитории Яндекса: создание сегментов, lookalike, использование в Директе и VK Ads.
Claude Code1 skill - yandex-360-admin2stars
Админ-API Яндекс 360: пользователи, домены, аудит, настройки организации.
Claude Code1 skill - clauterbox0stars
Flags overly verbose comments / JSDoc-TSDoc in TS/JS files using a configurable comment-to-subject length ratio, plus opt-in structural checks, and pushes back to Claude to fix them.
Claude Code - uroshcs-skills0stars
Uros's personal workflow skills. Includes 18 skills.
Claude Code18 skills - tools0stars
Werkzeuge mit ausschließlich expliziter Auslösung: /tools:ask und /tools:execute (die jeweils andere CLI non-interaktiv fragen bzw. beauftragen), /tools:unslop (KI-Muster aus Text entfernen), /tools:grimm (Sachverhalte ins Verwaltungs- und Anordnungsdeutsch überführen), /tools:orchestrator[-review|-cross] (Rollen-Prompt: steuern statt arbeiten, in drei Stufen bis zur anderen CLI als Arbeiter). Nachfolger der Plugins ask, unslop und grimm.
CodexClaude Code7 skills - philharmonie0stars
Projektaufträge mit Spec, getrenntem Generator und Evaluator sowie dauerhaftem Zustand bearbeiten; Cross-CLI-Fragen und begrenzte Schreibaufträge auch einzeln ausführen.
CodexClaude Code4 skills - ask0stars
[VERALTET – final, Nachfolger: tools] Cross-CLI-Zweitmeinung: fragt non-interaktiv die jeweils andere CLI (Codex aus Claude Code, Claude aus Codex) mit dem Flaggschiffmodell der installierten Version — /ask read-only (ohne Argument: Review des Projektstands), /execute mit Schreibrechten im Workspace. Der Host schreibt ein Handover, die andere CLI antwortet.
CodexClaude Code1 skill - daikenja0stars
Not an assistant -- a sage you consult. It reads the threads, documents, meetings and code you are buried in, asks what you are actually trying to do, and hands back something you can act on: a verdict on a draft, a review of how you handled a thread, a plain markdown ledger of what a project decided and what is still open. It never sends, never publishes, and never writes a word you did not ask for. The writing and remembering side of a working week.
Claude Code18 skills - claude-context-optimizer111stars
Claude 5-aware context optimization (Fable 5.1 / Opus 5 / Sonnet 5 pricing, 1h-cache-aware real costs & cache-break detection), SKILL.state-style task state that survives /compact, session baseline audit (/cco-overhead), auto .contextignore, self-calibrating estimates, delegation advisor, ru/en prompt coach, per-language token estimation, tunable thresholds via /cco-config, self-learning tool costs, team pattern sharing
Claude Code22 skills - stackwich1stars
Delivery discipline for AI coding agents. Writes the five-layer Stackwich policy (Prompt / Context / Harness / Loop / Graph) into CLAUDE.md and scaffolds advisor, executor and verifier subagents.
Claude Code1 skill - plan-staged-rollout1stars
Run big projects as many small sessions — not one huge one.
Claude Code1 skill - sbrodol-ai1stars
Ultra-verbose, deeply analytical, didactic and architectural communication mode (Anti-Caveman Protocol). Use when the user asks to "sbrodolare", wants maximum detail, a didactic walkthrough, an architectural justification, a paranoid security audit, or pure verbose entertainment.
Claude Code1 skill - rk0stars
Rk's Claude Code kit — skills, agents, hooks, rules for all repos
Claude Code50 skills - foreman0stars
Run Claude Code as a software team: architect, implementer, reviewer, and qa roles under a lead that reports to you.
Claude Code1 skill - health-harness0stars
Mindbowser Health Harness — an agent-agnostic discipline for building healthcare products with AI: the Build Loop + healthcare-compliance guardrails. Packaged as Claude Code skills.
Claude Code22 skills - workflow-suite1stars
Meta-bundle: the whole idea-to-shipped pipeline in one install — the core-suite baseline (candor, code-review, git-workflow, hindsight, secret-scanning, skill-router, stack-scan) plus taskmaster clarification and cards, task-runner execution with delegation contracts, approaches deliberation, code-architecture plan-before-code and verification, testing, debugging, and the ui-ux and security lanes the cards dispatch into. Replaces taskmaster-suite, process-suite and quality-principles-suite (2026-09-14). Fifteen members: over the host's 200k skill-listing floor by design — the README names the settings line. Uninstall cleanly with /workflow-suite:uninstall.
Claude Code - toolchain-experts1stars
Five expert reviewers that RUN the project's own static analysis instead of reciting a best-practice checklist: php-expert (phpstan, psalm, phpcs, rector), ts-expert (tsc, eslint, biome, oxlint), react-expert (the hook rules), vue-expert (vue-tsc, eslint-plugin-vue) and ui-expert (stylelint, pa11y, lighthouse-ci, axe). A detect-analyzers.sh mechanism reports which tool is configured, which one CI actually enforces, what the baseline already forgives and which rules are switched off — so the headline finding is the analyzer's blind spot, which no remembered rubric can produce. Read-only; fixes route to the existing workers.
Claude Code - overseer1stars
A program-level product owner for long, multi-session, multi-branch builds. /overseer:start "Build a CRM that manages clients" inventories the project and every installed plugin (user, project, local scope), asks one clarifying round, writes a milestone roadmap, then delivers each milestone on its own branch by briefing the taskmaster pipeline and task-runner execution with skill-pinned dispatch prompts. An accept script refuses to close a milestone until the app was driven in a real browser at three widths with a clean console — a green suite alone does not count. State under .claude/overseer/ lets /overseer:resume continue in a fresh session; a SessionStart hook announces it.
Claude Code - core-suite1stars
Meta-bundle: the user-scope baseline — the plugins worth enabling once, globally, and keeping on in every repo: secret-scanning's write-time block, candor's five-clause Stop gate and dormant terse mode, code-review's fan-in review with comment discipline, git-workflow's branch finish, hindsight's friction mining, skill-router's file-aware routing, and stack-scan's report, audit and project-tier scout. Replaces always-on-suite and quality-suite (2026-09-14). Membership rule: project-agnostic, safe as a permanent fixture, adds no interruption you did not ask for — why command-guard's ask tier is not here. Uninstall with /core-suite:uninstall. 200k windows: read the README first.
Claude Code - skills-equipe0stars
Les meilleures skills partagées de l'équipe : mode autonome à condition d'arrêt, qualité des réponses, parallélisation des gros lots, délégation à une IA moins chère. Toutes à activation manuelle. Une installation, des mises à jour pour tous.
Claude Code7 skills - pr0stars
Skills for working with pull requests — reviewing, describing, syncing, or otherwise assisting with the PR lifecycle.
Claude Code1 skill - oss0stars
Skills for maintaining and contributing to open source projects — sizing, filing, verifying, and fixing issues.
Claude Code4 skills - em-workflow0stars
Unified SDD + parallel-implementation + dynamic-review workflow. /em-workflow:develop drives spec → design (conditional; fully-autonomous DESIGN.md + HTML mockups + design-system/tokens.yaml — implementers never invent design, refinement happens post-implementation via /em-workflow:design) → plan (task decomposition with complexity/domain tagging) → fully-parallel implementation in git worktrees driven as a work queue (background implementer launches with up to max_parallel_implementers in flight, notification-driven slot refill, machine-written journal.jsonl + a fail-open guard-hook net (full hook inventory: em-workflow/references/implement-phase.md); each implementer merges its own task via flock-serialized merge-task.sh, resolving conflicts by parent-side adoption; before touching any file an implementer records which tests were already failing, so a post-implementation failure is attributed by set difference rather than re-investigated after the fact, and it commits that baseline plus the AC → test mapping with the observed red per criterion to test-docs/{feature}/{task_id}.tests.yaml alongside the implementation) → dynamic review selection (rule floor + orchestrator discretion): each selected perspective dispatches ONE non-Claude primary reviewer taken from the front of that perspective's registry-defined chain (Codex and, when the separately-installed vertex-review plugin is present, Vertex AI MaaS via LiteLLM), falling back to the Claude generic reviewer only when every chain entry for that perspective is unavailable; once every perspective's result is final, a single Opus evaluator subagent evaluates the whole round and the orchestrator decides the next action (auto-fix / another round / rework / completion) → bounded auto-fix → retrospect collection. Every workflow artifact (workflow.yaml, REQUIREMENTS.md/SPEC.md/DESIGN.md/IMPLEMENTATION.md/VERIFICATION.md, tasks/, feature-docs/{feature}/reviews/roundN.yaml, retrospect.yaml) lives exclusively inside a per-feature integration worktree/branch (em-workflow/{feature}/integration) — each update is committed there immediately via commit-docs.sh, so the main working tree stays clean until the final merge and a crashed/interrupted run resumes by enumerating integration branches rather than scanning the main tree. workflow.yaml-sourced shell commands are user-approved once per project (approval store outside the repo) and enforced deterministically by a bundled PreToolUse hook. Develop start (Step 0) runs a git-setup gate: it probes gitleaks availability and idempotently installs a gitleaks pre-commit hook, aborting the workflow when gitleaks is absent. License consistency is a cross-workflow constraint: create-spec records the project license (SPDX id) into workflow.yaml, create-plan checks new library choices against it (conflict → replace the library or relicense, user's choice), and the review phase adds a license perspective whenever the diff touches dependency manifests. Every phase's dialogue and file writes are separated by role: the orchestrator is the sole writer of workflow.yaml and the sole caller of AskUserQuestion, while a requirements-analyst investigates and a spec-writer authors REQUIREMENTS.md/SPEC.md for create-spec, an autonomous designer worker owns the design step, an implementation-planner proposes IMPLEMENTATION.md/tasks/VERIFICATION.md, and a rework-planner plans review/verify-triggered follow-up tasks — every worker returns a structured envelope (question packets or workflow patches) instead of writing shared state directly. A --batch flag runs develop unattended for externally-triggered headless launches: every AskUserQuestion gate resolves mechanically — a gate carried in a worker's question packet resolves per references/question-resolution.md's batch resolution sequence against references/batch-policies.yaml (spec ambiguities settled via a bounded Codex consultation loop, command approvals auto-recorded with refusal patterns still hard-failing), while a gate that never passes through a question packet (git-setup failure, feature selection, the review diff-size gate, capped auto-rework for review residuals and verify failures, keep-branch completion, and the rest) resolves per references/batch-mode.md's Non-packet gates table — batch never merges, and opens a PR only when the launch passes --pr: by default it removes the integration worktree and leaves the branch for the human to merge or PR from the main tree; failures still stop with a report. Also provides /em-workflow:review for standalone reviews, /em-workflow:retrospect for cross-feature lesson mining, /em-workflow:design for interactive post-implementation design refinement, /em-workflow:git-setup for standalone idempotent git local-config setup, and /em-workflow:gen-license for dependency-license analysis + LICENSE generation/relicensing.
Claude Code21 skills - em-review0stars
Standalone version of the em-workflow review phase. /em-review:multi-review reviews the current git diff (whole codebase when no diff) with two-layer dynamic perspective selection (rule floor + discretionary additions), skill-injected generic reviewers (Claude + conditional cross-model validation via GPT/Codex and, when the separately-installed vertex-review plugin is present, Vertex AI MaaS — security prefers Codex with a Vertex fallback on rate limit, other perspectives prefer Vertex with a Codex fallback when Vertex is not installed), cross-model agreement scoring, and bounded auto-fix (≤ 3 loops, skip with --report-only). Never commits; round records persist to /tmp/em-review/{repo-key}/ by default (--records <dir> to override) for nit-relitigation suppression across runs. Also reviews GitHub PRs by number/URL via gh pr diff (report-only).
Claude Code8 skills - worktree0stars
Parallel workspaces — create isolated git worktrees (plain repositories and submodules) so multiple agents can work on different plans simultaneously without file collisions.
Claude Code - session0stars
Session bookends — resume the previous session, and capture end-of-session state so the next resumes cleanly.
Claude Code - review0stars
Critical-review commands — /scrutinise for deep forensic review of recent work, /plan-review for multi-agent grounded review of a plan file.
Claude Code - requirements0stars
Requirements gathering — exhaustively interrogate a project's requirements before any code is written.
Claude Code