AI & ML

belentani7 Skill

Sistemas de contingencia de ejecucion sin internet con Ollama local.

belentani7 Skill

Routing inteligente de modelos: elige el LLM mas barato capaz de resolver la tarea y encadena fallbacks entre las APIs verificadas. Ver la skill `model-auto-router` para la politica completa.

belentani7 Bundle

Help a technical founder build whatever they want on Claude Managed Agents — an internal worker, a piece of their product, a customer-facing agent. Find out what they want to build, scope a v0, launch it in their account, grade it, iterate, and (if it should run on a clock) put it on a scheduled deployment, with everything bigger laid out as v1/v2. Use when a founder says "launch my agent", "/launch-your-agent", "build me a managed agent", or wants to build something on CMA. Keywords: managed agents, CMA, founder, launch, scheduled deployment, outcome rubric.

danicat Skill

Structural guide and editorial model for organizing technical articles, developer documentation, and READMEs. Applies the Inverted Pyramid model to place high-value summaries and actionable quickstarts first, cascading down to usage workflows, architectural details, and compliance. Activate when structuring technical articles, writing or refactoring README files, organizing documentation hierarchy, or making technical guides scannable.

belentani7 Skill

Curated registry of 100 essential AI agent repositories (2026) organized by category with asset types

belentani7 Bundle

Elige automaticamente el modelo/proveedor LLM mas barato capaz de resolver la tarea, con cadena de fallback entre las APIs verificadas del usuario. Trigger: 'que modelo uso', 'routing de modelos', 'abaratar tokens', 'fallback de proveedor', 'auto-router', 'modelo adecuado', 'optimiza coste LLM', o al arrancar tarea grande donde importe el coste.

belentani7 Skill

Connect OpenArt image and video generation models to AI tools (Claude, ChatGPT, Cursor) through the Model Context Protocol. Use when installing or authenticating the OpenArt CLI/MCP server, or connecting OpenArt models to an AI client.

gorock007 Bundle

Operate as Jarvis — the single main agent the user talks to in a Superset workspace, which plans end to end but delegates all building to worker agents on separate Superset terminals (`superset agents create` in the same workspace), picks the agent and model per task by what kind of decision the task contains (fable for undecided shape, opus/high for well-specified work, sonnet/haiku below that; Codex gpt-6-astra for any image/media asset or computer-use testing, gpt-5.6-sol for security; OpenCode's free models for tightly-fenced small and mechanical work, to spend no quota at all), drops fable for opus when Claude usage crosses 50% and moves the fleet to Codex past 75%, batches worker questions for the user, reviews handoffs, and alone owns git. Use whenever the user says "you're Jarvis", "act as the main agent", "spin off a worker", "delegate this", "what's open", "check on the workers", "merge the handoffs", "I'm running low on usage", or the repo has a CLAUDE.md that names Jarvis. Requires the Superset CL

gorock007 Bundle

Install the Jarvis workflow into the current repo — a main agent ("Jarvis") the user talks to, which delegates building to worker agents on separate Superset terminals, with usage-aware model routing (Claude Code fable/opus/sonnet/haiku, Codex astra/sol/terra/luna), a shared-checkout worker contract, and a handoff/OPEN.md system. Writes CLAUDE.md, AGENTS.md and handoffs/. Use whenever the user says "set up Jarvis", "add the Jarvis workflow", "make this repo use workers", "orchestrate with Superset terminals", or wants one coordinating agent plus parallel workers in Superset. Requires the Superset CLI (`superset`) — this workflow only works inside Superset.

erguzel Bundle

Vier-Augen (four-eyes) working agreement for agentic development inside a git repository - the agent works fast, the operator owns every permanent, destructive or publishing step. Load only when the operator invokes it by name ("vier-augen"). Never apply it on your own initiative.

wh1isper Bundle

Use email through the bundled mcp-email-server MCP server, diagnose bounded non-secret state, and hand account or credential setup to a user-operated CLI or authenticated local UI.

vybenetwork Bundle

Explore and call the Vybe Solana API catalog through MCP — discover endpoints, inspect OpenAPI params, then run authenticated live requests. Use when the user wants to browse APIs, find a specific route, understand parameters, or execute a Vybe HTTP call.

danicat Bundle

Inter-session communication mesh connecting independent terminal sessions on the same host via Unix Domain Sockets and dedicated Comms Subagents. Features sticky project identities, zero impersonation, non-blocking duplex messaging, crash-resilient mailbox spooling, and automatic clean restart lifecycles. Activate when coordinating multi-agent workflows across separate workspaces, delegating tasks between independent terminal sessions, relaying test failures or code diffs, or establishing a local agent mesh network.

danicat Skill

Radical simplification, red-teaming, and Occam's Razor devil's advocate for software architecture, PRDs, agent workflows, and technical designs. Challenges feature creep, speculative abstractions, and bloated specifications by proposing minimum viable primitives that deliver 90% of value with 10% of moving parts. Activate when reviewing complex technical proposals, pruning bloated architectures, red-teaming design docs, eliminating speculative features, or seeking the simplest possible path to production.

vybenetwork Bundle

Solana token and market intelligence via Vybe MCP — prices, holders, liquidity, candles, DEX trades, and top traders. Use when the user asks about a mint, token price, holders, markets, volume, OHLCV, or trader leaderboards.

danicat Bundle

Comprehensive guide to develop and improve Agent Skill performance. Contains best practices for skill formatting (frontmatter and metadata), naming, descriptions, fine-tuning activation triggers, evaluations, production-readiness and open sourcing. Activate when developing new skills or refining existing ones.

wrsilva Bundle

Builds native Android apps with the Android team's guides — Jetpack Compose UI, adaptive layouts, edge-to-edge, Compose styles, migrating XML views to Compose, Navigation 3 and predictive back; Android Gradle Plugin 9 upgrades, R8 keep rule analysis, profiling with Android Studio and Perfetto, and the Android CLI; Google Play policy compliance, Play Billing upgrades and the Engage SDK; Credential Manager restore and verified email; CameraX and Media3 Cast; App Functions and the ML Kit GenAI Prompt API; Wear OS, Android TV and display glasses. Use when writing, upgrading or auditing native Android (Kotlin) code for any of these topics, even if the user does not name the library.

allthingssmitty Bundle

Go type system advisor. Always use this skill when designing Go interfaces, working with structs and embedding, using generics, performing type assertions or type switches, defining custom error types, choosing between named types and aliases, or thinking about zero values. Use when the user says "how do I type this in Go", "when should I use an interface", "should I embed or compose", "how do I write a generic function", "how do I assert a type", "how do I wrap an error", "named type vs type alias", or "is my interface too big". Read this skill before answering any Go type system question.

allthingssmitty Bundle

TypeScript type system advisor. Always use this skill when designing TypeScript types, working with generics, narrowing union types, building discriminated unions, using utility types, or avoiding any. Use when the user says "how do I type this", "why is TypeScript complaining", "object is possibly null", "how do I narrow this", "how do I make this generic", "what's the right type for X", "how do I model this as a union type", or "how do I constrain a generic". Read this skill before answering any TypeScript type question.

tucan-marketing-digital Bundle

Workflows and best practices for using the HOT (Hub de Operaciones Tecnicas) MCP tools. Covers request lifecycle, milestone management, GitHub integration, checklists, and team coordination. Use when: creating requests via MCP, managing milestones, linking GitHub issues/branches/PRs, tracking progress with checklists, planning project workflows, or asking about HOT MCP tool usage.

odiumuniverse Skill

Resolve and prevent beadle sync conflicts. Use when `beadle status` or `beadle doctor` reports conflicts, when a sync or resolve is refused, or before editing agent config files that beadle synchronizes (rules, MCP servers, skills, permissions, memory).

allthingssmitty Bundle

Review or design REST and GraphQL APIs for correctness, consistency, and evolvability. Always use this skill when designing a new API, reviewing an existing one, choosing HTTP verbs, designing URL structure, defining error response shapes, thinking through versioning, or writing OpenAPI/Swagger specs. Use when the user says "review this API", "how should I model this resource", "is this endpoint design right", "should I use PUT or PATCH", "is this breaking", or "how should errors look". Use even when the question seems simple.

aitofy-dev Skill

Use when a coding agent must make a closed choice, noul (yes/no), or score judgment with TypeSafe Jev instead of guessing. Load it to proceed, ask, or stop; to route a skill; to guard a command; to verify a claim; to triage a record; to review a diff; or to pick the next legal action. Jev returns probabilities only. The host still writes the code and checks the result. Skip it for open-ended writing, or when one file or one command already answers the question.

vigneshwerv Skill

Use whenever the user asks to create a worktree for a task (e.g. "create a worktree for X", "spin up a worktree", "make a worktree for FRA-1234"). Creates the worktree via herdr, runs repo setup, splits three even panes (worker agent, Hunk review, user shell), and hands the agent the task context. Accepts `--agent <kind>` to choose the worker (claude, codex, gemini, ...); defaults to the agent running the skill.

uselemma Bundle

Drive the Lemma MCP from the terminal. Use when the user asks about Lemma issues, agent failures, incidents, or regressions; wants to review, validate, confirm, dismiss, fix, or resolve detections; or asks questions like "are there issues where users get frustrated", "any urgent lemma issues", "look through all urgent issues and validate each", or "dismiss everything under X impact." Also use when browsing, ranking, or acting on Lemma detections over MCP. Do not use for recording what an agent does in Lemma, or for writing agent context and artifacts from a codebase — that is lemma-artifacts.

uselemma Bundle

Read this codebase and record what its AI agents are for in Lemma. Use when the user asks to tell Lemma about their agent, set up or improve Lemma artifacts or agent context, record what an agent does, describe their agent to Lemma, bring Lemma's understanding up to date with the code, or fix a rule Lemma still has that the codebase no longer enforces. After its closing report it offers to hand off to lemma-tracing when the agent it settled on is not instrumented, whether or not anything was written. Do not use for installing tracing or fixing trace delivery and shape — that is lemma-tracing and lemma-diagnostics.

uselemma Bundle

Improve the quality of Lemma traces that are already landing. Use when the user asks whether traces are high quality, why a Lemma trace is missing tools, generations, or thread ids, to improve agent instrumentation, to audit the shape of traces already in Lemma, or whether traces match the high-quality traces guide. Do not use for first-time setup or missing delivery — that is lemma-tracing.

lucasgs520 Bundle

Use when building or editing any AI feature in n8n: AI Agents, Text Classifier, Information Extractor, Sentiment Analysis, Summarization Chain, Basic LLM Chain, embeddings, vector stores, single one-shot LLM calls, or AI media generation (image / audio / video) via the native LangChain provider nodes. Triggers on any `@n8n/n8n-nodes-langchain.*` node, "agent", "chat assistant", "LLM with tools", "tool calling", "fromAi", "system prompt", "memory window", "structured output", "outputParser", "function calling", "RAG", "vector store", "embeddings", "classify with AI", "extract fields with LLM", "sentiment analysis", "summarize with LLM", "single LLM call", chat triggers with files, AI image / video / audio generation, or any multi-turn or one-shot LLM behavior.

lucasgs520 Bundle

Decision guide for delegating to caveman-style subagents. Tells the main thread WHEN to spawn `cavecrew-investigator` (locate code), `cavecrew-builder` (1-2 file edit), or `cavecrew-reviewer` (diff review) instead of doing the work inline or using vanilla `Explore`. Subagent output is caveman-compressed so the tool-result injected back into main context is ~60% smaller — main context lasts longer across long sessions. Trigger: "delegate to subagent", "use cavecrew", "spawn investigator/builder/reviewer", "save context", "compressed agent output".

conceptual-machines Skill

Work with Tracktion Engine audio DAW framework. Use when creating/manipulating Edits, Tracks, Clips, Plugins, or Transport control. Covers the core model, playback, and plugin hosting patterns.

lucasgs520 Skill

Use when working with n8n workflows in any capacity. The always-on protocol for the n8n-skills plugin, loaded by the SessionStart hook every session. Routes to the right skill, summarizes every n8n MCP tool (closing the deferred-description gap), and lists the cross-cutting rules.

lucasgs520 Skill

Use when you want to expose an n8n workflow as a tool the coding agent can call. Two cases. (1) Wrap n8n API capabilities the MCP doesn't natively expose: folder CRUD, tag CRUD, instance metadata, credential creation. (2) Expose a general-purpose workflow as an agent tool: a workflow that calls a third-party API, runs business logic, or does any task you want the agent to invoke. Triggers on "expose as MCP tool", "build a tool for my agent", "I need to know X" where X isn't an MCP tool, "create folder", "create tag", or any capability gap.

lucasgs520 Bundle

Use when handling files, images, attachments, or binary data in n8n, OR when an AI agent needs to take a user-uploaded file as tool input or return a generated file. For Data Tables (schemas, dedup, persistent state), see the separate n8n-data-tables skill. Triggers on "file", "image", "PDF", "attachment", "binary", "upload", "download", chat trigger with files, agent tool that needs a file, vision/multimodal, or any handling of non-JSON file data.

lucasgs520 Bundle

Show real token usage and estimated savings for the current session. Reads directly from the Claude Code session log — no AI estimation. Triggers on /caveman-stats. Output is injected by the mode-tracker hook; the model itself does not compute the numbers.

agnuxo1 Skill

MASTER ROUTER — activates FIRST on every non-trivial task. Analyzes the problem, determines if a specialized tool/skill can resolve it without spending reasoning tokens, and orchestrates execution. Also enforces token-compression on ALL output. Trigger: ANY task involving computation, simulation, literature, verification, data, graphs, code, documents, or scientific analysis. No exceptions. The LLM is the orchestrator, not the executor.

agnuxo1 Bundle

Maximum token compression by substituting natural language with mathematical notation, code, and physical formulas. ALWAYS use this skill when Fran asks to reason, explain, analyze, calculate, model, design, compare, or optimize anything. Trigger words: "reason", "think", "explain", "analyze", "calculate", "model", "design", "how does X work", "compare", "optimize", "piensa", "razona", "explica", "analiza", "calcula", "modela", "diseña", "compara", "optimiza". Golden rule: if a natural language phrase has a compact formal equivalent → USE THE FORMAL.