paperclip-wiki-fetch

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

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime calls the Paperclip plugin bridge read routes (e.g., POST /api/plugins/paperclipai.plugin-llm-wiki/data/page-content), which return outsider-authored wiki page markdown/title that the agent then ingests into the LLM context.

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

• Potentially malicious external URL detected (high risk: 0.90). This skill makes runtime requests to the Paperclip plugin bridge (for example https://your-paperclip-host.example.com/api/plugins/paperclipai.plugin-llm-wiki/data/page-content) and injects the returned wiki markdown into the agent context, so the fetched external content can directly influence agent prompts/instructions.