smoke-test

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt tells the agent to ask the user for an LLM API key if not found and then create a .env file with that key (and suggests echoing env vars), which requires the LLM/agent to receive and embed secret values verbatim—an exfiltration risk.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly runs remote package scaffapers such as "npx create-mastra@" (and equivalent "yarn create mastra@", "pnpm create mastra@", "bunx create-mastra@"), which fetch and execute code from the npm registry at runtime and scaffold agent code/instructions that the smoke test relies on, so this external fetch both executes remote code and can directly control agent prompts.