quick-healthy-recipes

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

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime workflow can fetch arbitrary recipe pages from user-specified URLs via lightpanda fetch --dump markdown <url> and then ingest that fetched markdown text into the agent’s LLM context, which is outsider-authored free text (public web content).

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

• Potentially malicious external URL detected (high risk: 0.90). The skill uses the lightpanda CLI at runtime to fetch arbitrary external recipe pages (e.g., via "lightpanda fetch --dump markdown https://example-recipe-page"), which will pull remote content into the agent's context and directly inform prompts (and the setup also suggests running a remote installer via curl -fsSL https://get.lightpanda.io | sh, which executes fetched code), so these URLs present prompt-injection / remote-code risks.