Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add SantanderAI/ralph --skill maestrogit clone --depth 1 https://github.com/SantanderAI/ralphWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/santanderai/ralph/maestro)<a href="https://agentmods.dev/skills/santanderai/ralph/maestro"><img src="https://agentmods.dev/badge/skills/santanderai/ralph/maestro/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/santanderai/ralph/maestro"><img src="https://agentmods.dev/badge/skills/santanderai/ralph/maestro.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00080 | $0.01193 |
| Opus 5.5 | $0.00032 | $0.00477 |
| Sonnet 5.5 | $0.00016 | $0.00239 |
| Haiku 4.5 | $0.00008 | $0.00119 |
Grade A, and why
maestro scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 53 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Maestro
Curador del conocimiento local del proyecto sobre el que corre el bucle ralph. No implementa código de producto ni verifica evidencia de un bloque (eso es la juez): su trabajo es destilar, a partir de lo aprendido por el bucle, un conjunto de skills locales que reduzcan el groping de futuras iteraciones —tanto al planificar (fase 0) como al ejecutar— y eviten repetir los mismos errores.
Antes de actuar, lee primero las instrucciones locales aplicables del repositorio (AGENTS.md y, si existe, CLAUDE.md) y respétalas para comandos, entorno, evidencias y estilo. Si solo se está revisando o manteniendo esta skill (no hay un workspace con plan/ real sobre el que operar), no ejecutes las acciones del maestro.
Dónde viven las skills locales
Las skills locales que el maestro crea, extiende o borra viven en skills/, al mismo nivel que la carpeta plan/ del workspace donde corre el bucle (no bajo plan/, no en ~/.claude). Son locales al proyecto desde el que se invocan y viajan con su repositorio. Cada skill es un fichero skills/<nombre>/SKILL.md con la forma:
---
name: <nombre-en-kebab-case>
description: <una línea — para qué sirve y cuándo consultarla>
metadata:
tipo: metodologia | pitfall | decision
---
<el conocimiento reutilizable: la metodología a seguir, el pitfall a evitar (con su síntoma observable), o la decisión local tomada y su porqué. Conciso, accionable, una sola preocupación por fichero.>
Una skill local captura conocimiento durable y reutilizable del proyecto, no estado de la tarea ni notas de una iteración:
- metodología: la forma no obvia de hacer algo en este proyecto que costó varios intentos a ciegas descubrir (cómo se levanta el entorno, cómo se ejecutan los tests, el orden correcto de un proceso).
- pitfall: una trampa que el bucle pisó repetidamente, descrita con su síntoma observable y cómo evitarla.
- decision: una decisión local del proyecto (librería elegida, arquitectura, convención, restricción) que futuras iteraciones deben respetar o malgastarán iteraciones redescubriéndola.
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 1mo ago First seen · 53 lines · 80 tokens per session scan A a725a14b0936
maestro is a skill published in the GitHub repository SantanderAI/ralph (98 stars, last pushed 24d ago), licensed Apache-2.0. It adds 80 tokens to every session and 1,193 once invoked, about $0.0003 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
harness-creator
Build, audit, and improve harnesses that make AI coding agents reliable: AGENTS.md/CLAUDE.md instruction files, feature/state tracking, verification gates, scope boundaries, session handoff, memory persistence, context budgets, tool-permission safety, and multi-agent coordination. Use this whenever a coding agent is…
browserwing-executor
Control browser automation through HTTP API. Supports page navigation, element interaction (click, type, select), data extraction, accessibility snapshot analysis, screenshot, JavaScript execution, and batch operations.
memory-poisoning-detection
Detect persistent instruction injection or long-term memory poisoning. Focus on writing/retaining hostile instructions for future tasks, not data leakage.
dev-browser
Browser automation with persistent page state. Use when users ask to navigate websites, fill forms, take screenshots, extract web data, test web apps, or automate browser workflows. Trigger phrases include "go to [url]", "click on", "fill out the form", "take a screenshot", "scrape", "automate", "test the website"…
ask-user-question
Ask users questions via the UI. Use when you need clarification, user preferences, or confirmation before proceeding. The user CANNOT see CLI output - this tool is the ONLY way to communicate with them.
memos-memory-guide
Use the MemOS Local memory system to search and use the user's past conversations. Use this skill whenever the user refers to past chats, their own preferences or history, or when you need to answer from prior context. When auto-recall returns nothing (long or unclear user query), generate your own short search query…