Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Create structured technical specification documents that bridge product requirements and engineering implementation. Use when writing a tech spec, engineering spec, system design doc, or API specification. Produces a complete spec with problem statement, proposed solution, data model, API design, alternatives considered, security considerations, testing plan, and rollout strategy.
.claude/skills/mohitagw15856-technical-spec-template/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | 34% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -5% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 39% | 0% |
| case-18 | ✗→✓ | ▲ Improved | 43% | 0% |
| case-19 | ✗→✓ | ▲ Improved | 44% | 0% |
Escribe especificaciones técnicas que los ingenieros realmente lean — framing claro del problema, requisitos inequívocos, decisiones explícitas y compensaciones documentadas.
Pregunta al usuario por estos datos si no se proporcionan:
Escribe una especificación técnica cuando:
Sáltate la especificación para correcciones de bugs triviales o cambios de 1-2 horas.
Autor: Nombre] Estado: Borrador | En Revisión | Aprobado | Implementado Creado: Fecha] | Última Actualización: Fecha] Revisores: Líder de Ing., Arquitecto, PM, Seguridad si es necesario] PRD Relacionado: Enlace] | Epic de Jira: Enlace]
> 2–3 oraciones. ¿Qué problema estamos resolviendo y por qué ahora? Sin lenguaje de solución aquí.]
Objetivos (en alcance):
No-Objetivos (explícitamente fuera de alcance):
Cualquier trabajo previo, sistemas relacionados, o contexto que los ingenieros necesitan para entender el espacio de decisión. Enlaza a especificaciones previas, ADRs, o investigación.]
Enfoque de Alto Nivel: 2–4 oraciones describiendo la solución elegida. ¿Por qué este enfoque vs alternativas?]
Diagrama de Arquitectura del Sistema: Describe o incrusta: qué servicios están involucrados, cómo fluyen los datos, qué APIs se llaman]
Cambios del Modelo de Datos:
sql-- Nuevas tablas o cambios de esquema [Incluye DDL o definición de esquema]
Diseño de API:
[Endpoint] [Método]
Solicitud: { [campos y tipos] }
Respuesta: { [campos y tipos] }
Códigos de error: [lista]Detalles Clave de Implementación:
| Opción | Pros | Contras | Por Qué Rechazado | |---|---|---|---| | Alt 1] | Beneficios] | Desventajas] | Razón no elegida] | | Alt 2] | Beneficios] | Desventajas] | Razón no elegida] |
| Pregunta | Propietario | Fecha Vencimiento | Resolución | |---|---|---|---| | Pregunta sin resolver] | Nombre] | Fecha] | Pendiente] |
| Fase | Trabajo | Esfuerzo Estimado | |---|---|---| | Fase 1] | Qué se construye] | X días/points] | | Fase 2] | Qué se construye] | X días/points] | | Total | | X story points] |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 51,305 | 39,844 | -22% | 1 | 1 | 0% | 8,320 | 8,155 | -2% | 0 | 0 | — |
case-22 | pass→pass | 11,668 | 15,196 | +30% | 1 | 1 | 0% | 1,837 | 3,414 | +86% | 0 | 0 | — |
case-02 | fail→pass | 36,443 | 31,612 | -13% | 1 | 1 | 0% | 5,663 | 7,562 | +34% | 0 | 0 | — |
case-03 | fail→pass | 43,112 | 38,707 | -10% | 1 | 1 | 0% | 8,121 | 7,733 | -5% | 0 | 0 | — |
case-04 | pass→pass | 15,542 | 7,728 | -50% | 1 | 1 | 0% | 1,629 | 3,066 | +88% | 0 | 0 | — |
case-05 | pass→fail | 21,261 | 30,098 | +42% | 1 | 1 | 0% | 3,531 | 5,871 | +66% | 0 | 0 | — |
case-06 | pass→fail | 16,185 | 14,019 | -13% | 1 | 1 | 0% | 1,915 | 3,939 | +106% | 0 | 0 | — |
case-07 | pass→pass | 10,613 | 5,488 | -48% | 1 | 1 | 0% | 1,041 | 2,634 | +153% | 0 | 0 | — |
case-08 | pass→pass | 14,057 | 4,838 | -66% | 1 | 1 | 0% | 1,433 | 2,515 | +76% | 0 | 0 | — |
case-09 | pass→pass | 22,780 | 7,786 | -66% | 1 | 1 | 0% | 2,901 | 2,901 | 0% | 0 | 0 | — |
case-10 | fail→pass | 22,657 | 18,771 | -17% | 1 | 1 | 0% | 2,790 | 3,876 | +39% | 0 | 0 | — |
case-11 | pass→pass | 14,176 | 9,908 | -30% | 1 | 1 | 0% | 2,373 | 3,325 | +40% | 0 | 0 | — |
case-12 | pass→pass | 17,899 | 6,358 | -64% | 1 | 1 | 0% | 2,120 | 2,797 | +32% | 0 | 0 | — |
case-13 | pass→pass | 16,644 | 15,564 | -6% | 1 | 1 | 0% | 2,645 | 4,337 | +64% | 0 | 0 | — |
case-14 | pass→pass | 17,088 | 8,944 | -48% | 1 | 1 | 0% | 1,897 | 3,109 | +64% | 0 | 0 | — |
case-15 | pass→pass | 25,213 | 23,306 | -8% | 1 | 1 | 0% | 3,241 | 4,588 | +42% | 0 | 0 | — |
case-16 | pass→pass | 21,250 | 24,404 | +15% | 1 | 1 | 0% | 2,743 | 4,835 | +76% | 0 | 0 | — |
case-17 | pass→pass | 23,596 | 23,265 | -1% | 1 | 1 | 0% | 3,273 | 4,777 | +46% | 0 | 0 | — |
case-18 | fail→pass | 14,322 | 8,437 | -41% | 1 | 1 | 0% | 1,592 | 2,279 | +43% | 0 | 0 | — |
case-19 | fail→pass | 15,808 | 13,405 | -15% | 1 | 1 | 0% | 2,272 | 3,274 | +44% | 0 | 0 | — |
case-20 | pass→pass | 15,134 | 9,311 | -38% | 1 | 1 | 0% | 1,698 | 2,508 | +48% | 0 | 0 | — |
case-21 | fail→pass | 10,514 | 5,851 | -44% | 1 | 1 | 0% | 1,776 | 2,732 | +54% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +18 percentage points is the difference between those two pass rates over the 22 comparable cases. 2 cases got worse with the skill loaded, and they are included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.