site-launch-kit

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly asks the user to paste their Gemini API key into the chat and shows commands that insert that exact key into .env (and instructs echoing/using the key), which requires the LLM to receive and potentially output the secret verbatim.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The skill contains multiple deliberate patterns that risk credential exposure and supply‑chain compromise (silent third‑party MCP installation via npx, explicit prompt to paste API keys into chat and automated saving/reading of .env across parent directories, and instructions to avoid certain commands that look like evasion), so it should be treated as high risk for credential theft and supply‑chain/backdoor abuse.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 1.00). The skill explicitly runs runtime commands that fetch and execute remote packages (e.g., "claude mcp add nano-banana ... npx -y @zhibinyang/nano-banana-mcp" and installing/using agent-browser via "npm install -g agent-browser" as well as npx create-next-app / npx vercel / npx shadcn), which install and run external code from npm (examples: https://www.npmjs.com/package/@zhibinyang/nano-banana-mcp, https://www.npmjs.com/package/agent-browser, https://www.npmjs.com/package/vercel) during the skill runtime — these are runtime-executed external dependencies.