Saltar al contenido
TakoAPI
Volver a skills

shadow-ai-monitor

por techcodecrafter

Web & Frontend Development
290Descargas según clawskills.sh
1Vistas
0 me gustaInicia sesión para dar me gusta a esta skill
—Aún sin valoraciones
Tu valoración

Resumen

Shadow AI Monitor - Enterprise-grade dashboard for tracking employee AI tool usage, data exposure risks, and PIPEDA.

Descripción

Qué hace esta skill

Generates HTML dashboards for tracking employee AI tool usage across platforms like ChatGPT, Claude, and GitHub Copilot. Includes risk classification by data category, PIPEDA compliance scoring, and a demo data generator pre-loaded with realistic patterns for a 50-person organization. Runs entirely in the browser with no external dependencies.

Runs entirely locally with no external API calls or credentials, making it safe to use with sensitive organizational data in compliance-restricted environments.

Cuándo usarla

  • Presenting AI governance gaps to a company board
  • Auditing which AI tools employees use across a law firm
  • Checking PIPEDA compliance score before a regulatory review
  • Generating weekly shadow AI reports for a security team
  • Demoing AI risk dashboards to enterprise clients

Flujo de trabajo de ejemplo

ENTRADA

User asks: Generate a Shadow AI compliance dashboard for our organization

AGENTE
  1. 1

    Run generate_demo_data.js to create shadow-ai-data.json with employee AI usage records

  2. 2

    Run generate_dashboard.js with the generated JSON to produce shadow-ai-dashboard.html

  3. 3

    Open shadow-ai-dashboard.html to review the PIPEDA compliance score, risk breakdown, and employee heatmap

  4. 4

    Click individual employee bars to inspect their riskiest events and tools used

  5. 5

    Click Export to produce a board-ready PDF with letterhead styling

SALIDA

Interactive HTML dashboard showing a PIPEDA compliance score, color-coded risk event breakdown, and a downloadable PDF report

Instalación

clawhub install techcodecrafter/shadow-ai-monitor

Ejecuta el comando de la CLI o pega el prompt en Claude Code, Cursor, Codex o el agente de programación que uses.

Enlaces