linkfox-chuhaijiang-tiktok-video
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONPERSISTENCEEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from TikTok, such as video descriptions and user comments, which could contain malicious instructions.
- Ingestion points: Video captions and comments are retrieved via
chuhaijiang_video_search.pyandchuhaijiang_video_reviews.py. - Boundary markers: The skill provides display rules and limitations, but lacks explicit delimiters or warnings to ignore instructions embedded in fetched content.
- Capability inventory: The skill can write files to the local
linkfox/directory and perform network requests to vendor APIs. - Sanitization: No explicit sanitization of the remote TikTok content is performed.
- [PERSISTENCE]: The
references/onboarding.mdfile instructs users to persist API keys by modifying shell profile files (~/.zshrc,~/.bashrc). This is a standard credential persistence pattern for CLI-based tooling. - [EXTERNAL_DOWNLOADS]: The skill retrieves data and manages authentication via LinkFox vendor domains, including
tool-gateway.linkfox.com,api.linkfox.com, andagent-api.linkfox.com.
Audit Metadata