wechat-search
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill defines a research interface for searching the WeChat ecosystem via the SandBase API. No malicious code, hidden instructions, or suspicious behaviors were found.
- [INDIRECT_PROMPT_INJECTION]: The skill processes external search data, establishing a surface for indirect prompt injection. This is a common property of search-based skills and is mitigated by the lack of sensitive capabilities. 1. Ingestion points: Search results from wechat_search_v2_search and wechat_search_v2_search_videos (SKILL.md). 2. Boundary markers: The agent is instructed to cite sources and separate observations from analysis, though no technical markers are present. 3. Capability inventory: No file-write, shell execution, or network capabilities are present in SKILL.md or sandbase-api-map.md. 4. Sanitization: None identified.
Audit Metadata