
Product
Socket MCP Adds Org Alerts, Threat Feed Review, and Package Inspection
Socket MCP now lets AI assistants review org alerts, investigate threats using the Socket threat feed, and inspect package files in addition to dependency scoring.
secure-coding-agent
Advanced tools
Security-first AI agent workflow for subscription-based AI CLIs — Sonnet (Planner) · Haiku (Coder) · Gemini Pro (Reviewer) · Gemini Flash (Reporter) · Codex (Specialist).
Capa de orquestación multi-agente para suscripciones de AI — sin API keys.
Funciona sobre agent-security-policies, añadiendo el pipeline de roles encima de las reglas de seguridad.
Capa 1: npx agent-security-policies → AGENT_RULES.md + CLAUDE.md + AGENTS.md + skills de seguridad
Capa 2: npx secure-coding-agent → orquestación multi-agente + GEMINI.md + skills de pipeline
| Rol | Modelo | Suscripción | Archivo de instrucciones |
|---|---|---|---|
| 🧠 Planner | Claude Sonnet 4.6 | Claude Pro | CLAUDE.md (auto-cargado) |
| ⚡ Coder | Claude Haiku 4.5 | Claude Pro | CLAUDE.md (spawned) |
| 🔍 Reviewer | Gemini 3.1 Pro | Google One AI Premium | GEMINI.md (auto-cargado) |
| 📊 Reporter | Gemini Flash | Google One AI Premium | GEMINI.md (auto-cargado) |
| 🤖 Specialist | Codex o4-mini | ChatGPT Plus/Pro | AGENTS.md (auto-cargado) |
# Claude Code (Claude Pro)
npm i -g @anthropic-ai/claude-code
# Gemini CLI (Google One AI Premium)
npm i -g @google/gemini-cli
gemini auth login
# Codex CLI (ChatGPT Plus)
npm i -g @openai/codex
codex # → "Sign in with ChatGPT"
# Ejecutar sin instalar globalmente
npx secure-coding-agent --target /path/to/tu-proyecto
# O instalar globalmente
npm i -g secure-coding-agent
secure-coding-agent --target /path/to/tu-proyecto
git clone https://github.com/raomaster/secure-coding-agent.git
cd secure-coding-agent
# Instala ambas capas en el proyecto
./install.sh /path/to/tu-proyecto
# Con MCP servers (filesystem + memory)
./install.sh --mcp /path/to/tu-proyecto
# Solo capa de orquestación (si ya tienes agent-security-policies instalado)
./install.sh --no-security /path/to/tu-proyecto
Capa 1 (npx agent-security-policies --agent claude,codex,antigravity --skills):
AGENT_RULES.md — reglas de seguridad completas (OWASP ASVS 5.0, CWE Top 25, NIST SSDF)CLAUDE.md — instrucciones de seguridad para ClaudeAGENTS.md — instrucciones de seguridad para Codex.agent/rules/security.md — reglas para Gemini (formato antigravity).claude/commands/ — 7 skills de seguridad: /sast-scan, /secrets-scan, /dependency-scan, /container-scan, /iac-scan, /threat-model, /fix-findingspolicies/ — YAML: owasp_asvs.yaml, cwe_top25.yaml, llm_security.yaml, owasp_masvs.yamlCapa 2 (este repo):
CLAUDE.md ← append: protocolo de orquestación multi-agenteGEMINI.md — rol de reviewer/reporter (no existe en capa 1).claude/commands/plan.md — pipeline: Fase 1+2 research + plan.claude/commands/code.md — pipeline: Fase 3 delegar a Haiku.claude/commands/review.md — pipeline: Fase 4 revisión Gemini Pro.claude/commands/report.md — reporte ejecutivo Gemini Flash.claude/commands/full-cycle.md — pipeline completo end-to-endTu petición en Claude Code
↓
/full-cycle "implementar X"
↓
Sonnet 4.6 — Planner
├── Explora codebase (Glob/Grep/Read)
├── Hace preguntas clarificadoras
└── Crea plan de tareas atómicas
↓
[confirmación del usuario]
↓
Haiku 4.5 × N — Coder (workers paralelos)
└── CLAUDECODE= claude --model haiku --print ...
↓
Gemini 3.1 Pro — Reviewer
└── cat archivos | gemini -m pro --yolo -p "security review..."
↓
Gemini Flash — Reporter
└── echo hallazgos | gemini -m flash --yolo -p "executive report..."
GEMINI.md + AGENT_RULES.md se cachean| Comando | Descripción |
|---|---|
/plan | Research del codebase + plan estructurado de tareas |
/code | Delegar implementación a Haiku worker(s) |
/review | Security review con Gemini 3.1 Pro |
/report | Reporte ejecutivo con Gemini Flash |
/full-cycle | Pipeline completo plan→code→review→report |
| Comando | Herramienta |
|---|---|
/sast-scan | Semgrep — vulnerabilidades CWE en código |
/secrets-scan | Gitleaks — credentials hardcodeadas |
/dependency-scan | Trivy fs — CVEs en dependencias |
/container-scan | Trivy image — CVEs en Docker |
/iac-scan | KICS — misconfigs Terraform/K8s/etc. |
/threat-model | Gemini Pro — STRIDE threat modeling |
/fix-findings | Remediación de hallazgos |
secure-coding-agent/
├── CLAUDE.md # Append: protocolo de orquestación multi-agente
├── GEMINI.md # Rol reviewer/reporter para Gemini CLI
├── install.sh # Instala ambas capas en cualquier proyecto
├── .claude/
│ ├── settings.json # MCP: filesystem + memory (con --mcp)
│ └── commands/ # Solo skills de pipeline (seguridad → agent-security-policies)
│ ├── plan.md
│ ├── code.md
│ ├── review.md
│ ├── report.md
│ └── full-cycle.md
└── README.md
Con --mcp, se instala .claude/settings.json con:
Tanto Claude Code como Gemini CLI son clientes MCP — puedes agregar servidores MCP adicionales para compartir contexto entre agentes.
FAQs
Security-first orchestration for Claude Code and OpenCode-based AI workflows.
The npm package secure-coding-agent receives a total of 17 weekly downloads. As such, secure-coding-agent popularity was classified as not popular.
We found that secure-coding-agent demonstrated a healthy version release cadence and project activity because the last version was released less than a year ago. It has 1 open source maintainer collaborating on the project.
Did you know?

Socket for GitHub automatically highlights issues in each pull request and monitors the health of all your open source dependencies. Discover the contents of your packages and block harmful activity before you install or update your dependencies.

Product
Socket MCP now lets AI assistants review org alerts, investigate threats using the Socket threat feed, and inspect package files in addition to dependency scoring.

Product
Socket Firewall blocks malicious VS Code and Open VSX extensions before install, protecting developers from compromised editor marketplaces.

Research
More than 140 Mastra npm packages were compromised in a supply chain attack that used a typosquatted dependency to deliver a cross-platform infostealer during installation.