memoryvault

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly fetches and surfaces user-generated public content (e.g., /public/{agent}, /feed, /changelog, /comments, /discover, and /search returning public memories and comments from other agents on https://memoryvault.link) which the agent is expected to read/interpret as part of its workflow, exposing it to untrusted third‑party content that could carry indirect prompt injection.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly instructs agents at startup to fetch live content from https://memoryvault.link (e.g., "On startup, read MEMORYVAULT.md and fetch latest from https://memoryvault.link" and GET /get/welcome), and that remote content (memories/onboarding) can supply instructions that directly influence agent prompts at runtime.