juez

juez is a skill for Claude Code, Codex from SantanderAI/ralph. It costs 77 tokens per session (1,031 once invoked), scanned A, original, Apache-2.0.

A Spanish-language review skill for plan and task Markdown files. It checks planned work at defined checkpoints and can review overall progress in read-only mode.

In plain words
What is it for?
It is for reviewing implementation plans, judging completed subtasks, suggesting plan changes and auditing progress without changing product code.
Why use it?
It provides an independent check on whether task evidence is reproducible and helps identify or unblock work that has been rejected repeatedly.

Skill for Claude CodeCodex

Install

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.

agentmods
npx agentmods add skills/santanderai/ralph/juez
Any agent
npx skills add SantanderAI/ralph --skill juez
Clone the repo
git clone --depth 1 https://github.com/SantanderAI/ralph

Made for: Claude Code, Codex.

Wrote 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.

agentmods badge for juez

README.md
[![agentmods](https://agentmods.dev/badge/skills/santanderai/ralph/juez.svg)](https://agentmods.dev/skills/santanderai/ralph/juez)
Your own site
<a href="https://agentmods.dev/skills/santanderai/ralph/juez"><img src="https://agentmods.dev/badge/skills/santanderai/ralph/juez.svg" alt="Measured on agentmods" height="20"></a>
Per session 77 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,031 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00077 $0.01031
Opus 5 $0.00039 $0.00515
Sonnet 5 $0.00015 $0.00206
Haiku 4.5 $0.00008 $0.00103

Measured 5d ago against content hash d02311f13052, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

juez 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 5d 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.

skills/juez/SKILL.md · 36 lines

What it actually says

Juez

Revisor externo e independiente sobre plan/task/*.md. No implementa código de producto.

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 ejecutes las acciones de juez sobre una task concreta.

Ubicación del plan: por defecto el plan vive en plan/plan.md y sus tareas en plan/task/*.md desde la raíz del repositorio. Si el usuario nombra explícitamente otra ruta (otro directorio, otro nombre de fichero), usa la que indique en lugar de la por defecto. Todas las referencias a estas rutas en las acciones de esta skill son relativas a esa convención.

Tiene cuatro acciones distintas. Antes de actuar, decide cuál corresponde:

Situación Acción Cargar
El usuario invoca explícitamente /juez auditar Auditar el enfoque global del bucle y proponer cambios al plan (solo lectura, informe en conversación) auditar.md
Hay una subtarea [juez desatascar] pendiente [ ] (insertada al acumularse rechazos en un bloque) Desatascar el bloque: leer plan, logs (.ralph/logs, vía subagentes) y cómo valida la juez; si el criterio es inalcanzable, documentar problema+solución e insertar una tarea de implementación del arreglo desatascar.md
Hay subtareas de implementación marcadas [x] con un [juez] posterior pendiente Juzgar la evidencia del primer bloque pendiente (uno por invocación) y desmarcar si falla juzgar.md
No hay subtareas [x] pendientes de juzgar; se quiere validar el plan (creación o revisión) Planear: insertar los [juez] que falten (y los [crear evidencia] necesarios antes de ellos) y editar el plan para que ninguna subtarea exija intervención del usuario planear.md

Auditar solo se activa por invocación explícita del usuario; nunca entra en el ruteo automático de apertura de ficheros. Cuando se pide explícitamente, tiene prioridad sobre las demás.

Entre las acciones automáticas, la prioridad es: desatascar (si hay un [juez desatascar] pendiente) > juzgar (si hay evidencia que juzgar) > planear (cobertura de checkpoints). El desatascado se antepone porque se inserta justo antes de un [juez] que lleva rechazos acumulados: hay que diagnosticar el atasco antes de seguir juzgando.

Reglas comunes a todas las acciones:

  • No implementa código. Si detecta carencias, las refleja como feedback, como nuevos checkpoints, o como un bloque de problema+solución con su tarea de implementación; nunca corrigiéndolas tú.
  • No asume. "Parece correcto" no basta; sin evidencia recolectable, el resultado es no evaluable, no positivo.
  • Determinismo. La evidencia que dependa de orden, tiempo, red o entorno exige contexto fijado (semilla, mock, fixture, snapshot).
  • Respeta el entorno del repo. No uses flujos legacy si las instrucciones locales indican otro camino. Los artefactos desechables que crees al vuelo para comprobar pueden vivir en /tmp; los arneses persistentes de [crear evidencia] se commitean en el repo (nunca en /tmp).
  • Una línea en la task. El feedback que se deja en una subtarea es una única línea en lenguaje natural; nunca incluye fragmentos de código, rutas, comandos ni stack traces. La conversación sí puede resumir brevemente qué se hizo y qué evidencia falló o pasó. (La excepción es el bloque de diagnóstico que escribe desatascar.md, que sí es un bloque estructurado y detallado en el cuerpo del fichero de tarea.)

En uso normal, carga solo el fichero correspondiente (juzgar.md, planear.md, desatascar.md o auditar.md) según la tabla anterior y sigue sus instrucciones. Para revisar o actualizar la propia skill, puedes leer varios.

Files

What ships with it

4 files 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.

Changes

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.

  1. 5d ago First seen · 36 lines · 77 tokens per session scan A d02311f13052

Subscribe to this mod's changes

juez is a skill published in the GitHub repository SantanderAI/ralph (89 stars, last pushed 3d ago), licensed Apache-2.0. It adds 77 tokens to every session and 1,031 once invoked, about $0.0004 per session on Opus 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.

Related

Other skills, from other repositories

edgeone skill scanner

Scan any agent skill for security risks before you install or use it. Powered by Tencent Zhuque Lab A.I.G (AI-Infra-Guard). 100% local static analysis — no file contents or credentials leave your device. Compatible with CodeBuddy, Cursor, Windsurf, Claude Code, OpenClaw and more. Triggers on: 这个 skill 安全吗, skill 安全扫描…

Tencent/AI-Infra-Guard · 148 tokens

baby-sit

Monitor a GitHub pull request until CI is green, diagnose failures, and rerun only evidence-backed flaky GitHub Actions jobs.

langchain-ai/open-swe · 30 tokens

continual-learning

Nightly refinement of an existing per-repo review-style prompt using this reviewer's own finding outcomes. Read confirmed (resolved-by-commit / thumbs-up) and dismissed (thumbs-down) findings, promote the bug patterns the team actually fixes, demote the false-positive patterns, reconcile against the current prompt…

langchain-ai/open-swe · 89 tokens

nano-banana-pro-openrouter

Deterministic OpenRouter image generation adapter for Nano Banana Pro / Gemini image models. Use as skillexec when a meta-skill needs local image files and structured IMAGEREADY records without spawning an LLM agent.

opensquilla/opensquilla · 49 tokens

skill-creator-linter

Internal tool (not user-invocable). Called by meta-skill-creator as a DAG step (kind: agent) to lint a candidate meta-skill SKILL.md against G1 (parse + reference check + xmlescape grep + structural lint) and G2 (scheduler dry-run with stub executors). Deterministic, sub-second, no LLM. Returns JSON diagnostics.

opensquilla/opensquilla · 84 tokens

paper-abstract-author

Write the abstract after the paper body has been revised, using the final claims and evidence.

opensquilla/opensquilla · 23 tokens