pipeline-judge-dual

pipeline-judge-dual is a skill for Claude Code, Codex from danielgap/openclaw-planitor. It costs 88 tokens per session (3,620 once invoked), scanned A, original, MIT.

A two-reviewer quality check for a business plan. One reviewer focuses on facts, calculations and consistency, while the other looks at strategy and creativity; the results are combined into one report.

In plain words
What is it for?
Use it after the final plan and investor evidence are prepared to check facts, financial information, completeness and internal consistency. It produces separate reviews, a combined report and feedback for a rerun.
Why use it?
It reduces the chance that a plan contains unsupported claims, calculation errors or missing sections. The report also records problems and feedback for another revision.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one.

Good fit Use it after the final plan and investor evidence are prepared to check facts, financial information, completeness and internal consistency. It produces separate reviews, a combined report and feedback for a rerun.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/danielgap/openclaw-planitor/pipeline-judge-dual
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.

Any agent
npx skills add danielgap/openclaw-planitor --skill pipeline-judge-dual
Clone the repo
git clone --depth 1 https://github.com/danielgap/openclaw-planitor

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 pipeline-judge-dual

README.md
[![agentmods](https://agentmods.dev/badge/skills/danielgap/openclaw-planitor/pipeline-judge-dual/github.svg)](https://agentmods.dev/skills/danielgap/openclaw-planitor/pipeline-judge-dual)
Your own site
<a href="https://agentmods.dev/skills/danielgap/openclaw-planitor/pipeline-judge-dual"><img src="https://agentmods.dev/badge/skills/danielgap/openclaw-planitor/pipeline-judge-dual/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.

agentmods 80×15 button for pipeline-judge-dual

Your own site · 80×15
<a href="https://agentmods.dev/skills/danielgap/openclaw-planitor/pipeline-judge-dual"><img src="https://agentmods.dev/badge/skills/danielgap/openclaw-planitor/pipeline-judge-dual.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 88 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 3,620 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
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 5.1 $0.00088 $0.03620
Opus 5 $0.00044 $0.01810
Sonnet 5 $0.00018 $0.00724
Haiku 4.5 $0.00009 $0.00362

Measured 11d ago against content hash a64621448021, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-11, from the pricing page.

Security

Grade A, and why

pipeline-judge-dual 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 11d 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/pipeline-judge-dual/SKILL.md · 313 lines

How it starts

The opening of the file, as written. The whole thing — 313 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Pipeline — FASE 9: Judge Dual-LLM

Dos LLMs evalúan el plan desde perspectivas complementarias. Planitor sintetiza ambas evaluaciones y produce un informe de calidad unificado con validación fiscal, scoring de completitud y feedback documentado.

Cuándo se ejecuta

Después de FASE 8 (Investor Pack). Cuando plan-final.md e investor-evidence-pack.md están generados.

Input

Fuente Ruta
Plan final projects/{proyecto}/plan-final.md
Todos los artefactos projects/{proyecto}/v{n}/
Research web projects/{proyecto}/research-web.json
Research benchmarks projects/{proyecto}/research-benchmarks.json

Output

Producto Ruta
Evaluación Juez A projects/{proyecto}/v{n}/judge/JUDGE-A.json
Evaluación Juez B projects/{proyecto}/v{n}/judge/JUDGE-B.json
Informe final sintetizado projects/{proyecto}/v{n}/JUDGE-REPORT.json
Feedback para re-run projects/{proyecto}/v{n}/judge/FEEDBACK-FOR-RERUN.json

Los Dos Jueces

Agente Modelo Enfoque
modelo-a modelo-principal Analítico-riguroso — fact-check numérico, coherencia interna, verificación de cálculos, consistencia de datos entre secciones, validación contra artefactos fuente
modelo-b modelo-secundario Estratégico-creativo — viabilidad del modelo, diferenciación real, fuerza de la propuesta, calidad narrativa, persuasión del documento, honestidad estratégica

Ventaja dual-LLM: Un modelo detecta errores y inconsistencias mecánicas; el otro evalúa si el plan realmente convencería a un inversor/banco. La síntesis produce una evaluación completa y equilibrada.

Paso 1: Launch Paralelo

ejecutar como sub-tarea(id=modelo-a, task:
  "Lee skills/pipeline-judge-dual/SKILL.md para entender el formato.
   Lee plan-final.md en projects/{proyecto}/.
   Lee TODOS los artefactos en projects/{proyecto}/v{n}/.
   Lee research-web.json y research-benchmarks.json en projects/{proyecto}/.
   
   Tu ENFOQUE: analítico-riguroso.
   
   Ejecuta fact-check exhaustivo:
   1. NÚMEROS: ¿Coinciden break-even, márgenes, cash flow entre secciones y artefactos?
   2. COHERENCIA INTERNA: ¿Buyer persona consistente? ¿Pricing igual en todas las secciones?
   3. DATOS vs FUENTES: ¿Los datos del plan coinciden con GROUND-TRUTH y research?
   4. SENSIBILIDAD: ¿3 escenarios presentes? ¿Variable make-or-break identificada?
   5. REDUNDANCIA: ¿Datos repetidos? ¿Secciones duplicadas?
   6. COMPETENCIA: ¿URLs verificadas? ¿Mínimo 3 competidores?
   7. ESTACIONALIDAD: ¿Basada en datos INE? ¿Zona emisora/receptora correcta?
   8. DATOS ANTIGUOS: ¿Hay datos >6 meses sin marcar ⚠️?
   9. UNIT ECONOMICS: ¿LTV/CAC ≥ 3:1 en escenario base? ¿Techo de facturación?
   10. CURVA DE INGRESOS: ¿Realista para el sector? ¿Efecto novedad modelado?
   11. FISCALIDAD — IVA: ¿Se desglosa IVA repercutido (21%) vs IVA soportado? ¿Se contempla IVA soportado en inversiones?
   12. FISCALIDAD — IRPF: ¿Se estima IRPF por tramos sobre beneficios? ¿Se incluyen retenciones?
   13. FISCALIDAD — CUOTA AUTÓNOMOS: ¿Se incluye cuota RETA mensual (~300€/mes mínimo)?
   14. FISCALIDAD — RÉGIMEN FISCAL: ¿Se indica régimen fiscal (módulos vs estimación directa)? ¿Es coherente con el modelo?
   15. FISCALIDAD — DEDUCCIONES: ¿Se identifican deducciones fiscales aplicables (inversión en equipo, I+D, creación empleo)?
   16. COMPLETITUD: ¿El plan tiene TODAS las secciones obligatorias? Verificar checklist:
       - Resumen ejecutivo
       - Análisis de mercado (TAM/SAM/SOM)
       - Competencia con URLs
       - Modelo de negocio (9 bloques Canvas)
       - Plan financiero 3 años
       - Análisis de sensibilidad
       - Plan operativo
       - Riesgos y mitigación
       - Plan de financiación
       - Timeline con hitos
       - Análisis fiscal (IVA, IRPF, cuota autónomos, régimen)
       - Plan regulatorio / marco legal
       - Plan de salida/liquidación
   
   Para cada issue: clasificar como crítico/warning/sugerencia.
   Scores 1-10 por criterio.
   
   Genera un JUDGE-REPORT.json.
   Escríbelo en projects/{proyecto}/v{n}/judge/JUDGE-A.json"
)

ejecutar como sub-tarea(id=modelo-b, task:
  "Lee skills/pipeline-judge-dual/SKILL.md para entender el formato.
   Lee plan-final.md en projects/{proyecto}/.
   Lee TODOS los artefactos en projects/{proyecto}/v{n}/.
   
   Tu ENFOQUE: estratégico-creativo.
   
   Evalúa como un inversor/banco leería el plan:
   1. PROPUESTA DE VALOR: ¿Es realmente diferencial o 'me too'? ¿Convence?
   2. MODELO DE NEGOCIO: ¿Es sostenible y escalable? ¿O es un trabajo disfrazado?
   3. NARRATIVA: ¿El plan engancha? ¿Se lee de tirón? ¿O es aburrido y genérico?
   4. HONESTIDAD: ¿Hay optimismo infundado? ¿Se ocultan riesgos? ¿Edulcoración?
   5. DIFERENCIACIÓN: ¿La ventaja competitiva es defendible? ¿O fácil de copiar?
   6. VIABILIDAD ESTRATÉGICA: ¿El GTM tiene sentido? ¿Los canales son realistas?
   7. PERSUASIÓN: ¿Convince a un inversor escéptico? ¿Y a un banco conservador?
   8. CALIDAD DE CONSULTORÍA: ¿Se aplicaron frameworks de expertos? ¿O solo se nombraron?
   9. PROPÓSITO: ¿Resuelve un problema real? ¿O es una solución buscando problema?
   10. FACTOR X: ¿Tiene algo que lo hace memorable y diferente de otros planes del sector?
   
   Para cada issue: clasificar como crítico/warning/sugerencia.
   Scores 1-10 por criterio.
   
   Genera un JUDGE-REPORT.json.
   Escríbelo en projects/{proyecto}/v{n}/judge/JUDGE-B.json"
)

Read the full file on GitHub · 313 lines

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. 11d ago First seen · 313 lines · 88 tokens per session scan A a64621448021

Subscribe to this mod's changes

pipeline-judge-dual is a skill published in the GitHub repository danielgap/openclaw-planitor (5 stars, last pushed 4mo ago), licensed MIT. It adds 88 tokens to every session and 3,620 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 2026-08-31.

Related

Other skills, from other repositories

systematic-debugging

Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.

obra/superpowers · 21 tokens

local-ai-agents

Build local-first AI agents that run entirely on a developer workstation with Microsoft Foundry Local and Qwen function-calling models. Covers Small Language Models (SLMs), the OpenAI-compatible local endpoint, sandboxed local tools, local RAG with Chroma, local MCP servers, hybrid cloud/local routing, and the…

microsoft/ai-agents-for-beginners · 200 tokens

next-cache-components-adoption

Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…

vercel/next.js · 95 tokens

insight-error-page

Write or audit an insight-kind error page for the Next.js dev overlay. Use when creating a new errors/ .mdx page, auditing an existing one, or checking that a page matches the framework fix cards. Covers page structure, title alignment, FixCard cards with Copy prompt button, code snippets, terminology verification…

vercel/next.js · 83 tokens

next-cache-components-optimizer

Drive a Next.js route to instant navigation by setting up an agentic loop, under Cache Components / PPR, on initial load (hard navigation) and client-side navigation (soft navigation). Encode the goal as a failing @next/playwright instant() e2e and work it to green, one verified route at a time; the shipped test then…

vercel/next.js · 170 tokens

next-partial-prefetching-adoption

Turn on Partial Prefetching in a Next.js app and work through the insights it surfaces. Use when the user wants to enable or adopt Partial Prefetching, flip the partialPrefetching flag, opt routes in with export const prefetch = 'partial', audit Link prefetch={true} behavior, preserve existing prefetched UI with…

vercel/next.js · 103 tokens