菜单
Usar cuando se necesita auditar la seguridad de un proyecto con pipeline Red Team / Blue Team.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Guide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or services, whether in Python (FastMCP) or Node/TypeScript (MCP SDK).
Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.
Detecta el Bus Factor por modulo en un repositorio git usando el algoritmo CST(change-size-ratio). Genera JSON con BF, owners, riesgo, y avisa cuando un solo dev conoce un modulo critico.
Genera CONTEXT_DOME.md para modulos con Bus Factor bajo. Captura conocimiento tacito: proposito, decisiones no obvias, dependencias, runbook minimo, knowledge owners y plan de distribucion.
Usar cuando se quiere ejecutar tareas de bajo riesgo de forma autónoma durante la noche.
Mapear la superficie de ataque de un dominio: subdominios, OSINT, typosquatting.
| name | adversarial-security |
| description | Usar cuando se necesita auditar la seguridad de un proyecto con pipeline Red Team / Blue Team. |
| summary | Pipeline Red Team + Blue Team + Auditor independiente. Scoring CVSS, mapeo STRIDE, gap analysis. Output: informe con score 0-100 y recomendaciones. |
| maturity | stable |
| context | fork |
| context_cost | medium |
| agent | security-attacker |
| category | governance |
| tags | ["security","adversarial","red-team","blue-team"] |
| priority | high |
| trigger | {"type":"keyword","keywords":["vulnerabilidad","pentest","sql injection","xss","inyeccion","security audit","red team","blue team"]} |
If you were dispatched as a subagent to execute a specific delegated task, skip this skill's full orchestration workflow. Execute only the assigned task, report result (DONE / DONE_WITH_CONCERNS / BLOCKED), and return. This guard prevents runaway skill activation in nested agent contexts.
Doble opt-in (SPEC-186): → si no: ❌ ABORT
bash scripts/savia-double-optin-check.sh \
--skill adversarial-security --confirm-autonomous
Requiere AMBOS: ADVERSARIAL_SECURITY_ENABLED=true Y flag explicito.
CVSS simplificado para proyectos internos:
| Factor | Peso | Valores |
|---|---|---|
| Attack Vector | 0.3 | Network (1.0), Adjacent (0.7), Local (0.5), Physical (0.2) |
| Complexity | 0.2 | Low (1.0), High (0.5) |
| Privileges | 0.2 | None (1.0), Low (0.6), High (0.3) |
| Impact | 0.3 | High (1.0), Medium (0.6), Low (0.3) |
score = sum(factor × peso) × 10 → escala 0-10
| Categoría | Pregunta clave | Controles típicos |
|---|---|---|
| Spoofing | ¿Puedo suplantar a otro? | Auth, MFA, tokens |
| Tampering | ¿Puedo modificar datos? | Integridad, signing, HMAC |
| Repudiation | ¿Puedo negar una acción? | Audit logs, timestamps |
| Info Disclosure | ¿Puedo acceder a datos? | Encryption, access control |
| DoS | ¿Puedo tumbar el servicio? | Rate limiting, WAF |
| Elevation | ¿Puedo escalar privilegios? | RBAC, least privilege |
# npm: audit de dependencias
npm audit --json 2>/dev/null | jq '.vulnerabilities | length'
# pip: safety check
pip-audit --format=json 2>/dev/null
# dotnet: audit
dotnet list package --vulnerable --format json 2>/dev/null
score = 100 - (critical×25 + high×10 + medium×3 + low×1) Cada fix verificado recupera los puntos. Floor: 0.