alexferrari88/skills

Última actividad de origen registrada: 18 de octubre de 2026 a las 05:37

Catálogo de SkillsMP actualizado: 29 de octubre de 2026 a las 18:30

skills recopiladas: 10

Estrellas en GitHub: 1
Forks en GitHub: 0

Instalar con un asistente de IA

Copia este prompt en el asistente de IA que estés usando.

Follow https://skillsmp.com/skill-install/prompt.md to install Agent Skills from https://github.com/alexferrari88/skills.

README del repositorio

Archivo de origen del README: README.md

Leer el README original

Mostrar README completo

# skills A public collection repo for reusable Hermes and skills-compatible agent skills. ## Included skills - **`hermes-deep-research`** — Hermes-native deep research skill for multi-source synthesis, verification, parallel subresearch, and decision-grade reporting. - **`dspy-hermes-runtime-auth`** — reuse Hermes Agent's active runtime provider, base URL, and credentials inside DSPy without copying API keys. - **`dspy-rlm-hermes-runtime-auth`** — companion skill for deciding when `dspy.RLM` is the right abstraction and wiring it to Hermes runtime auth cleanly. - **`tasker`** — automating Android tasks, profiles, scenes, and actions using Tasker configurations, JavaScriptlets, intents, and shell commands. - **`teach`** — evidence-backed tutoring runtime for stateful lessons, practice, feedback, transfer, and spaced review. - **`agy-cli`** — use Antigravity CLI (`agy`) for agent sessions, non-interactive prompts, model discovery, and plugin management. - **`local-web-extract`** — extract complete URL content through self-hosted Firecrawl with deterministic Crawl4AI fallback. - **`socratic-argument-tutor`** — examine, defend, or stress-test a concrete argument through neutral, evidence-aware guided dialogue. - **`high-efficiency-learning-design`** — design, audit, or troubleshoot high-efficiency mastery-based learning, prerequisite sequencing, practice, and retention in hierarchical domains. - **`ambitious-problem-solving`** — elevate tasks, prompts, plans, and workflows across altitude (outcomes and autonomous loops), depth (raw substrates, expert tools, and outside judges), and range (cross-disciplinary methods without role-play). ## Install Use this collection repo with any skills-compatible agent: ```bash npx skills add https://github.com/alexferrari88/skills --skill <skill-name> ``` Examples: ```bash npx skills add https://github.com/alexferrari88/skills --skill hermes-deep-research npx skills add https://github.com/alexferrari88/skills --skill dspy-hermes-runtime-auth npx skills add https://github.com/alexferrari88/skills --skill dspy-rlm-hermes-runtime-auth npx skills add https://github.com/alexferrari88/skills --skill tasker npx skills add https://github.com/alexferrari88/skills --skill teach npx skills add https://github.com/alexferrari88/skills --skill agy-cli npx skills add https://github.com/alexferrari88/skills --skill local-web-extract npx skills add https://github.com/alexferrari88/skills --skill socratic-argument-tutor npx skills add https://github.com/alexferrari88/skills --skill high-efficiency-learning-design npx skills add https://github.com/alexferrari88/skills --skill ambitious-problem-solving ``` ## Skill summaries ### `hermes-deep-research` Best for evidence-led deep research where important claims need full-source reading, cross-checking, and structured output. ### `dspy-hermes-runtime-auth` Best for portable DSPy code that should follow Hermes runtime changes across providers, models, and custom endpoints without duplicating secrets. ### `dspy-rlm-hermes-runtime-auth` Best for long-context tasks where the model should recursively inspect and compute over a large context object instead of forcing one giant prompt through a single context window. ### `tasker` Best for creating, modifying, or debugging Android automation profiles, tasks, scenes, custom JavaScriptlets, shell commands, and intents in Tasker. ### `teach` Best for adaptive tutoring sessions where the learner should attempt, receive feedback, retry, transfer, and leave durable learning records instead of just reading an explanation. ### `agy-cli` Best for running and managing Antigravity CLI (`agy`) sessions, models, plugins, and non-interactive prompts from an agent workflow. ### `local-web-extract` Best for complete URL extraction when ordinary agent fetching fails or is incomplete, using self-hosted Firecrawl first and Crawl4AI as a visible fallback. ### `socratic-argument-tutor` Best for learning argument analysis by mapping a concrete claim, its grounds, and its warrant, then testing one disputed reasoning step at a time. ### `high-efficiency-learning-design` Best for designing, auditing, or troubleshooting mastery-based learning sequences, practice dose, retrieval spacing, interleaving, and diagnostic remediation in hierarchical technical domains. ### `ambitious-problem-solving` Best for upgrading under-scoped prompts, plans, analyses, and workflows across altitude (decisions and continuous shadow loops over one-off artifacts), depth (raw data, executable expert instruments, and external validation oracles), and range (noun-to-verb functional reframing and cross-disciplinary methods without theatrical role-play). ## License MIT

Cargando el README del repositorio…

Publicaciones

Ver todos →

Alex

@thealexferrari18 oct 2026 07:09 UTC

Anuncio de un SkillPensamiento crítico

Resumen de la publicación · inglés

ambitious-problem-solving pushes a task past literal compliance by raising ambition, deepening evidence and adversarial validation, and borrowing methods across disciplines.

Publicación original

Ver detalles del Skill · skills/ambitious-problem-solving

Skills en este repositorio

Desarrolladores de software Profesores postsecundarios, todos los demás

2 categorías ocupacionales · 80% clasificado

skill

ocupación

descripción

actualizado

local-web-extract

sin clasificar

Extract full content from webpages, articles, PDFs, and JavaScript-heavy URLs through self-hosted Firecrawl with automatic Crawl4AI fallback. Use when ordinary built-in web fetching fails, is incomplete, or when the user asks for full page extraction. Do not…