gzh-subscribe
Pass
Audited by Gen Agent Trust Hub on Aug 11, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is well-structured and follows security best practices. It uses environment variables for credential storage and restricts network communications to the official vendor domain (doubaoya.com). No evidence of malicious behavior or obfuscation was found.
- [PROMPT_INJECTION]: The skill possesses an inherent attack surface for indirect prompt injection because it ingests and displays article titles and metadata from external WeChat accounts.
- Ingestion points: Data retrieved from the API response in
scripts/account_works.pyand displayed viaSKILL.md. - Boundary markers: Absent; data from the API is directly interpolated into Markdown tables.
- Capability inventory: Execution of local scripts via
python3as defined in theSKILL.mdworkflow. - Sanitization: The instructions do not specify any validation or sanitization for the article titles retrieved from external sources.
Audit Metadata