AI & ML

diegopherlt Bundle

Crea archivos markdown de sub-agentes de Claude Code (con YAML frontmatter y system prompt) listos para colocar en ~/.claude/agents/, .claude/agents/ o agents/ de un plugin. Usa esta skill proactivamente cuando el usuario diga "crea un agente", "create an agent", "necesito un sub-agente que...", "scaffold an agent", "haz un auditor", "make a code reviewer agent", "agente que revise", "agente que audite", "agente que investigue", o cuando describa funcionalidad especializada que conviene aislar en un sub-agente (ej. "necesito algo que revise PRs", "quiero un agente que analice seguridad"). Cubre seleccion de tools con principio de menor privilegio, tuning de description para auto-invocacion, eleccion de arquetipo (auditor / researcher / implementer / orchestrator), calibracion de modelo + effort segun la carga de razonamiento de la tarea (mas mecanico = modelo mas ligero y menos effort) e incluye un sistema de scoring de confianza 0-100 con umbral >=80 para agentes tipo auditor.

lossyrob Skill

Reference documentation for PAW multi-phase implementation workflows. Provides activity tables, artifact structure, stage guidance, and PR routing patterns. Workflow enforcement rules are in PAW.agent.md.

lossyrob Skill

Pre-PR review activity skill for PAW workflow. Reviews implementation against spec before Final PR creation with configurable single-model, multi-model, or society-of-thought execution.

lossyrob Skill

Interactive pre-spec ideation utility skill. Agent-led Q&A to progressively clarify vague ideas, research codebase context, and produce structured WorkShaping.md artifact.

harperreed Skill

Use only for a current operative request that explicitly invokes $roborev-fix, or a direct Agent Hook instruction; do not invoke from literal syntax in quoted, pasted, or historical text

getty104 Skill

Pencil CLI（`pencil`コマンド）だけを使って既存の.penファイル（Pencilで作成されたデザインファイル）をAIプロンプトで修正・更新するスキル。ユーザーが.penファイルの編集、ボタン追加、レイアウト変更、UIデザインの調整、Pencilデザインの更新などを依頼した場合に必ずこのスキルを使用する。エージェントモード（`pencil --in --out --prompt`）で同一パスを指定して既存ファイルを上書き編集し、編集後はインタラクティブモード（`pencil interactive`）でNodeツリーから「**編集したコンポーネントのNodeだけ**」を特定して`get_screenshot` / `export_nodes` でPNG出力し、`.pen`と同階層の`snapshots/`ディレクトリに保存する。Pencil MCPには依存せず、`pencil` コマンドのみで完結する。

getty104 Skill

Pencil CLI（`pencil`コマンド）だけを使って、.penファイル（Pencilで作成されたデザインファイル）の中のNodeのデザインデータ（属性・構造）とスクリーンショット画像を読み取り専用で取得するスキル。Node IDが分かっているケースだけでなく、名前の正規表現（例: 「ヘッダー」「.*Button」）、Nodeタイプ（frame / text / image など）、再利用可能コンポーネント、特定フレーム配下、ドキュメント全体のトップレベルなど、**ID以外の指定方法**にも対応する。ユーザーが「.penのこのNodeの中身を見せて」「特定コンポーネントのデザインデータを取り出して」「Nodeのスクリーンショットだけ欲しい」「ヘッダーの構造を確認したい」「ボタンのスタイルをコピーしたい」「再利用可能コンポーネント一覧を見せて」「ドキュメント全体の構造を覗きたい」「全てのテキストNodeを取得して」のように.pen内の要素の調査・参照・確認・抜き出しを依頼した場合に必ずこのスキルを使う。インタラクティブモード（`pencil interactive`）で `batch_get` の `nodeIds` / `patterns` / `parentId` を使い分けてNode属性をJSONで取得し、`get_screenshot` / `export_nodes` で画像を`.pen`と同階層の`snapshots/`にPNG出力する。編集はしない（`save()`を呼ばない）ため、対象ファイルは絶対に書き換わらない。Pencil MCPには依存せず`pencil`コマンドのみで完結。

stbenjam Bundle

Use when a deeper level of code review is requested. Multi-agent panel code review with specialist reviewers and forced runtime reproducers for all BLOCKING bug findings. Optionally posts to GitHub/GitLab as a PENDING review.

lossyrob Bundle

Guide for creating effective skills. This skill should be used when users want to create a new skill (or update an existing skill) that extends an agent's capabilities with specialized knowledge, workflows, or tool integrations.

lossyrob Skill

Holistic review of planning artifacts bundle (Spec.md + ImplementationPlan.md + CodeResearch.md) with configurable single-model, multi-model, or society-of-thought execution. Catches cross-artifact consistency issues before implementation begins.

lossyrob Skill

Reviews PRs or code changes using multiple AI models, synthesizes findings, and interactively applies fixes. Uses GPT 5.2, Gemini 3 Pro, and Opus 4.5 for diverse perspectives.

lennetech Skill

Optimizes this Claude Code marketplace using specialized sub-agents. Each agent is an expert for one element type (skills, commands, agents, hooks, mcp) and loads only relevant documentation. A marketplace agent validates cross-references and latest features. Triggers on "optimize marketplace", "sync with best practices", or when user wants to improve plugin quality.

harperreed Skill

Access meme-rodeo MCP tools for searching, browsing, and analyzing memes. Use when the user asks about memes, wants to search for memes, get meme stats, or interact with the meme-rodeo platform.

harperreed Skill

Use when the user asks to check on agents, manage their fleet, babysit panes, see what needs attention, get a status report, or when they say "check on my agents", "what needs action", "fleet status", "babysit", "brief me". Requires the control MCP server to be registered and the daemon running.

harperreed Skill

Use when evaluating, designing, or pressure-testing the business model of an AI agent product. Triggers on "agent business model", "agent economics", "agent canvas", "evaluate agent product", "agent pricing", "agent unit economics", "agentic business", "AI agent company", "agent cost structure".

artwist-polyakov Bundle

Управление сессиями AI агентов через agent-deck CLI. Триггеры (RU): "запусти агента", "запусти саб-агента", "создай сессию", "проверь сессию", "проверь статус", "покажи вывод агента", "что агент ответил". Triggers (EN): "launch sub-agent", "create sub-agent", "start session", "check session", "show agent output".

lennetech Bundle

Migration guides, release notes, and error solutions for updating @lenne.tech/nest-server in npm mode. Covers version-specific breaking changes, stepwise major upgrades, and starter comparisons; delegates execution to the lt-dev:nest-server-updater agent. Activates on nest-server version upgrades, "pnpm run update", TypeScript errors after an upgrade, or stepwise migration planning. NOT for vendored cores (use nest-server-core-vendoring). NOT for writing NestJS code (use generating-nest-servers). NOT for general npm updates (use maintaining-npm-packages).

lennetech Skill

Decides per project whether `.claude/agent-memory/**` is committed, records that decision next to the memory it governs, and curates the notes before every commit that carries them. Activates before committing whenever agent-memory files are dirty: git:ship, ticket-cycle, review runs, or any standalone commit.

lennetech Bundle

Coordination patterns and worktree isolation for parallel operations this session starts: Agent Teams (independent sessions with messaging) and parallel subagent spawning (Agent tool with isolation worktree). Covers when teams beat single agents and what they cost in tokens. Activates on "agent team", "parallel review", "batch rebase", when a command evaluates team suitability via CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS, or when spawning several file-modifying subagents at once. NOT for single sequential subagents. NOT for sessions the user started themselves (use coordinating-peer-sessions).

jovermier Skill

Use this agent when performing security audits, vulnerability assessments, or security reviews of code. Triggers on requests like "security review", "check for vulnerabilities", "OWASP compliance check".

jovermier Skill

Use this agent when conducting thorough research on a repository's structure, documentation, and patterns. Specializes in analyzing architecture files, examining GitHub issues, reviewing contribution guidelines, and searching for implementation patterns. Triggers on requests like "analyze repository", "research codebase structure", "find patterns in repo".

jovermier Skill

Use this agent when reviewing database migrations, data models, or any code that manipulates persistent data. Specializes in validating referential integrity, transaction boundaries, and data validation rules. Triggers on requests like "data integrity review", "database safety check".

0x0pointer Bundle

AI/LLM red-team assessment using OWASP LLM Top 10 (2025), the OWASP AI Testing Guide (AITG v1, Nov 2025), and OWASP MCP Top 10 runtime testing for agentic/MCP targets. Tests prompt injection, jailbreaks, system prompt leakage, sensitive data extraction, excessive agency, improper output handling, model extraction, content bias, evasion, membership inference, MCP token exposure, and MCP command injection. Combines three tools: FuzzyAI (single-turn jailbreak fuzzing), Garak (probe-based scanning), and promptfoo (plugin-based red-team eval, incl. multi-turn jailbreak/crescendo strategies) - each covering different OWASP categories. Includes a conditional MCP recon phase and a post-access AI infrastructure phase (chained from /post-exploit). Produces an OWASP LLM Top 10 + AITG + MCP coverage matrix, findings per category, an architecture diagram, and PoCs. Chains into /gh-export.

0x0pointer Bundle

Deep web exploitation beyond initial scanning. Covers SQLi (blind, OOB, second-order), NoSQL injection, GraphQL injection (introspection, batching, mutation abuse), XSS (reflected/stored/DOM with source-sink analysis), SSTI (Jinja2/Twig/Freemarker/ERB RCE), SSRF chains, file upload bypass (polyglots), XXE (blind, DOCX/SVG, Content-Type switching), deserialization (Java/PHP/Python/.NET), command injection, path traversal / LFI wrapper bypasses, race conditions, CSRF, JWT attacks (none/key confusion/kid injection), HTTP request smuggling (CL.TE/TE.CL/H2), CRLF injection, open redirect chains, CORS exploitation, web cache deception/poisoning, OAuth misconfiguration, prototype pollution, session management, and business logic flaws. Uses sqlmap, commix, xsser, wapiti, davtest, and manual http payloads - every technique includes real payloads and code. Chains from /pentester or /api-security, into /post-exploit on RCE, and into /ai-redteam when an LLM/AI endpoint is found.

artwist-polyakov Bundle

Поиск и ресёрч через Perplexity API: сырая выдача (Search API), ответ с цитатами (Agent API), deep research в фоне, чтение страниц. Cache-first: крупные результаты уходят в cache/ и читаются грепом. Triggers: perplexity, перплексити, perplexity search, sonar api, найди в интернете, поищи в сети, web search, свежая информация, deep research, глубокое исследование, ответ с источниками, прочитай страницу, fetch url, что пишут про.

lennetech Skill

Rules for working alongside other Claude Code sessions the user started in parallel. Defines the four coordination channels (Linear and Git, the ledger, `ListAgents`, `SendMessage`), the seven message occasions (LANDED, CLAIM, CONFLICT, SOLVED, READY, ASK, ORIGIN), and the permission boundary a peer message can never cross. Carries change provenance: how to tell what this session wrote from what it found, via `change-provenance.sh`, and how to get the intent behind a foreign change from the session that wrote it. Activates on "peer session", "andere Session", "wer arbeitet gerade woran", "wer hat das geändert", "fremde Änderungen", when `ListAgents` shows live peers, when a message arrives, when a review or ship meets uncommitted work nobody in this session made, or when a change affects a repo a peer consumes. NOT for agent teams this session spawns (use coordinating-agent-teams). NOT for subagents (use the Agent tool directly).

lennetech Skill

Knowledge base for projects that vendored the @lenne.tech/nest-server core into their source tree (projects/api/src/core/, no npm dependency). Covers detecting a vendored project, the one-way vendor model, when `src/core/` may be touched, the flatten-fix pattern, upstream sync and upstream-PR workflows, and cosmetic-vs-substantial classification. Activates on vendored core discussions, "sync core from upstream", or vendor-sync conflicts. NOT for npm-based nest-server updates (use nest-server-updating). NOT for writing new NestJS code (use generating-nest-servers).

lennetech Skill

Final browser validation after implementation AND review have succeeded. Boots the app via `lt dev up`, seeds realistic `@test.com` data, derives a step-by-step test list from the diff (every affected page, role, flow, empty/error state, mobile pass, console + network sweep), then walks it autonomously via Chrome DevTools MCP. Fixes everything it finds, including pre-existing issues, in the same loop. Activates as the last step of any ship-oriented workflow, and on "im Browser prüfen", "durchklicken". NOT a substitute for implementation, code review, or automated E2E tests, which run before.

lennetech Skill

The lenne.tech fullstack production-readiness checklist. Defines the eight pillars (configuration and secrets, observability, health and lifecycle, security hardening, data durability, resilience under load, deployment hygiene, runbook and rollback) with file/line evidence, severity classification, and a machine-parseable report block. Activates on "production ready", "produktionsreif", "Go-Live", "kann das live", or when a release must be gated before shipping. NOT for OWASP code-level security review (use the security-reviewer agent). NOT for npm dependency audits (use maintaining-npm-packages).

0x0pointer Skill

Deep API security assessment beyond surface scanning, covering the full OWASP API Security Top 10 (2023): Broken Object Level Authorization (BOLA/IDOR), Broken Authentication, Broken Object Property Level Authorization (mass assignment + excessive data exposure), Unrestricted Resource Consumption, Broken Function Level Authorization (BFLA), Unrestricted Access to Sensitive Business Flows, SSRF via API parameters, Security Misconfiguration, Improper Inventory Management (shadow/zombie/deprecated endpoints, v1/v2 drift), and Unsafe Consumption of third-party APIs. Works across REST, GraphQL, gRPC, SOAP, and MCP servers. Discovers APIs from OpenAPI/Swagger, GraphQL introspection, gRPC reflection, .well-known, and traffic capture. Uses kiterunner, ffuf, schemathesis, openapi-fuzzer, graphql-cop, inql, jwt_tool, and mitmproxy with real payloads. Chains from /pentester or /codebase, into /web-exploit for injection points, /post-exploit on RCE, and /ai-redteam when an LLM/AI endpoint is found.

0x0pointer Skill

Full penetration test using MCP tools — recon, scanning, exploitation, and reporting. Tailored for OpenCode (BYO LLM). Supports network/web targets and local codebases. Chains into analyze-cve, threat-modeling, and remediate skills automatically.

rajbos Skill

Describes the data files available in the coding agent environment after copilot-setup-steps runs. Use when analyzing downloaded session logs or aggregated usage data.

rajbos Skill

Refresh token estimator and model pricing JSON files with latest data from AI model providers

rajbos Skill

Analyzing GitHub Copilot session log files to extract token usage, model information, and interaction data. Use when working with session files, understanding the extension's log analysis methods, or debugging token tracking issues.

rajbos Skill

Analyze coverage of the vscode-extension's tool-family definitions (DEFAULT_TOOL_FAMILIES in vscode-extension/src/toolFamilies.ts) against the canonical tool-name list in src/toolNames.json and/or a live agent tool list, and suggest which family missing tool names likely belong to. Use when the Tool Output Token Analysis tab shows too many tools bucketed under "Other Tools", after adding a new editor adapter, after a new agent/CLI's tool list grows, or when asked to improve/expand the preconfigured tool families.

claude-world-director-mode-lite Skill

Systematic debugging method: 5-step root-cause analysis (capture, isolate, hypothesize, investigate, fix & verify) plus common bug-pattern reference. Use when errors, exceptions, test failures, or unexpected behavior appear. Loaded automatically by the debugger agent.