playlet-wechat-feed
Pass
Audited by Gen Agent Trust Hub on Aug 11, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill invokes a local Python script,
scripts/fetch_playlet_feed.py, to perform API calls. This script is used to fetch and return JSON data to the agent for processing. - [EXTERNAL_DOWNLOADS]: The script performs network requests to the vendor's official API endpoint at
doubaoya.comto retrieve short drama metadata. This is a functional requirement and follows standard vendor resource access patterns. - [PROMPT_INJECTION]: The skill processes untrusted data (article titles and links) obtained from an external API, creating an attack surface for indirect prompt injection during the summarization and clustering phase.
- Ingestion points: External data enters the agent context via the
fetch_playlet_feed.pyscript output, which contains article metadata from WeChat. - Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present in the skill when the agent is asked to cluster or summarize the articles.
- Capability inventory: The skill is limited to reading data and outputting text; it does not possess high-risk capabilities such as arbitrary file writing, privilege escalation, or persistence mechanisms.
- Sanitization: Article titles and links are structured into Markdown tables, but no specific validation or sanitization of the content is performed before the agent summarizes the data.
Audit Metadata