qa-studio

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly instructs the agent to accept and include inline secret values in a shell command (--secret KEY=VALUE) and to have users paste credentials into the chat (which the agent will execute), forcing the LLM to receive and output secrets verbatim into the chat, model context, and shell command line.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.85). The required runtime workflow for “ready-to-run import” (SKILL.md:25 and building-tests.md:185-197) can take user-supplied free-form credentials via chat and inject them into the agent/LLM context and shell command line (explicitly warned as landing in the chat transcript, model context, and command line).