supermemory

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

• Third-party content exposure detected (high risk: 0.90). The skill's documentation and API reference explicitly show ingestion of arbitrary public URLs and web pages (e.g., POST /v3/documents 'content' can be a URL in references/api-reference.md and quickstart/examples in SKILL.md/references/quickstart.md), and those ingested third-party contents are retrieved via profile()/search and directly used to build prompts and drive agent actions—meeting all criteria for exposure to untrusted, user-provided content that could enable indirect prompt injection.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill makes runtime calls to the Supermemory API (e.g., https://api.supermemory.ai and its endpoints like POST /v4/search and /v3/documents) and then directly injects the returned profile/memories into system prompts (e.g., building systemMessage with response.profile and response.memories), so externally-fetched content is used at runtime to control agent prompts.