wechat-moments
Pass
Audited by Gen Agent Trust Hub on Aug 7, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructions direct the user to clone an external repository (https://github.com/OpenClaudia/wechat-moments) to obtain the scripts necessary for the skill's operation. This is documented as the 'plumbing' required to handle the WeChat desktop client and database export.- [PROMPT_INJECTION]: The skill is designed to process and summarize untrusted data from WeChat Moments (朋友圈), which constitutes an indirect prompt injection attack surface where content within posts could attempt to influence the agent's behavior.
- Ingestion points: The skill reads a plaintext database produced by an external WeChat export tool.
- Boundary markers: Absent. There are no explicit delimiters defined to separate the untrusted post content from the agent's instructions, though the skill instructs the agent to summarize rather than repeat content verbatim.
- Capability inventory: The skill documentation states that reading and writing files is the remit of the scripts in the referenced repository.
- Sanitization: Not specified. The instructions do not define methods for escaping or filtering potential malicious instructions embedded in the feed content.
Audit Metadata