gzh-ai-feed

Pass

Audited by Gen Agent Trust Hub on Aug 11, 2026

Risk Level: SAFEDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [DATA_EXFILTRATION]: The skill performs network operations to an external domain not on the predefined whitelist to fetch article data.
  • Evidence: The script scripts/fetch_ai_feed.py sends POST requests to https://doubaoya.com/api/apis/gongzhonghao/gongzhonghao-ai-feed/call containing the user's API key in the authorization header.
  • Context: This network activity is the primary purpose of the skill and is documented as the method for retrieving WeChat feed data.
  • [PROMPT_INJECTION]: The skill processes untrusted content from external sources, which could be used for indirect prompt injection attacks.
  • Ingestion points: External article titles and metadata are fetched via scripts/fetch_ai_feed.py and passed to the agent's context as JSON.
  • Boundary markers: The instructions in SKILL.md lack explicit delimiters or instructions for the agent to ignore potentially malicious content embedded within the fetched article titles.
  • Capability inventory: The agent is instructed to cluster the articles and generate insights, a process that involves analyzing the untrusted string content.
  • Sanitization: No sanitization or filtering is performed on the external data before it is rendered into Markdown tables for the agent to process.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 11, 2026, 07:48 PM
Security Audit — agent-trust-hub — gzh-ai-feed