Skillfrisk

Security-scan any AI-agent skill or MCP server before installing or publishing it. Use when the user asks to check, audit, or vet a skill/MCP server, or before adding third-party skills from GitHub. Detects prompt injection, secret exfiltration, remote code execution, destructive commands, and dangerous MCP permissions.

Topicspot Updated
skillfrisk scanning a malicious skill and a clean one

File contents

Topicspot/skillfrisk commit 2ab8ae3947

Frequently asked questions

npx get-skill add topicspot/skillfrisk