dalle

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly instructs the agent to ask the user to paste their Pixazo API key into the conversation and to save that exact key to ~/.pixazo/api-key, so the LLM must receive and handle the secret verbatim (high exfiltration risk).

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

• Potentially malicious external URL detected (high risk: 1.00). The skill explicitly instructs the agent to "Fetch: https://www.pixazo.ai/models/dalle.md" at runtime to load exact parameter names and schemas, which means remote content would be fetched and used to control prompt/request construction.