subconscious-dev

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

• Third-party content exposure detected (high risk: 0.90). The skill's tools system (see SKILL.md and references/tools-guide.md) includes platform/function tools like "web_search" and "webpage_understanding" that fetch public web pages and search results and Subconscious injects those tool responses into the agent's reasoning/context, so untrusted third‑party content can directly influence tool use and decisions.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly defines runtime "function" tools that Subconscious will POST to user-hosted endpoints (e.g., https://your-server.com/search or ngrok URLs like https://your-ngrok-url.ngrok.io) and injects the returned JSON into the agent's context/ reasoning, so those external URLs (https://your-server.com/search) can directly influence prompts/agent behavior at runtime.