douyin-subscribe
Pass
Audited by Gen Agent Trust Hub on Jun 12, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill makes network requests to
https://redfox.hk/story/api/dyData/searchWorkListto retrieve Douyin account and video data. This domain belongs to the skill's author (redfox-data) and is used for its primary functionality. - [COMMAND_EXECUTION]: Instructs the agent to use
automation_updatefor managing recurring tasks. It specifies that account IDs should be passed directly as command-line arguments (--accounts) in these automated tasks, which involves the agent constructing and updating shell commands. - [DATA_EXFILTRATION]: User-provided identifiers (Douyin IDs) are sent to the external service
redfox.hkto perform data lookups. This is a functional requirement of the skill. - [PROMPT_INJECTION]: The skill includes instructions for a 'Silent Execution Principle' (静默执行原则), directing the agent to perform validation and subscription steps without displaying intermediate output to the user. While this reduces transparency into the agent's actions, it appears intended for user experience rather than concealment of malicious activity.
- [CREDENTIALS_UNSAFE]: The skill requires an API key (
REDFOX_API_KEY) but correctly instructs users to manage it via environment variables or configuration files rather than hardcoding it into the skill's logic.
Audit Metadata