social-sbti

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The skill prompt instructs the agent to ask the user for authentication credentials (auth_token and ct0 cookies for X) and then pass them directly as command-line arguments to a configuration script (--auth-token and --ct0), exposing sensitive secrets through the LLM's execution context.

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

• Third-party content exposure detected (low risk: 0.10). The workflow fetches public social media posts from Jike or X based on a user-supplied target URL or handle, which are then read by the LLM for personality scoring, exposing the agent to outsider-authored free text.

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

• Potentially malicious external URL detected (medium risk: 0.30). The templates/card.html file imports runtime client-side script from an external CDN (https://esm.sh/html-to-image@1.11.13), constituting an unverified third-party runtime external dependency via a public registry transport.