aliyun-dashvector-search
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill installs the
dashvectorPython package as part of its setup process. - Evidence:
SKILL.mdcontains the instructionpython -m pip install dashvector. This is the official SDK for Alibaba Cloud's DashVector service. - [INDIRECT_PROMPT_INJECTION]: The skill processes data retrieved from an external vector database, which serves as a potential vector for indirect prompt injection if the retrieved content contains malicious instructions.
- Ingestion points: Untrusted data enters the agent's context through query results in
SKILL.mdand thecollection.querycall inscripts/quickstart.py. - Boundary markers: The skill does not define specific delimiters or instructions to the agent to treat the retrieved content as data rather than instructions.
- Capability inventory: The skill includes capabilities for creating collections, upserting documents, and querying data. It does not include high-privilege system operations like arbitrary command execution or file system writing beyond the validation scripts.
- Sanitization: There is no evidence of sanitization or filtering applied to the document fields retrieved from the vector database before they are processed by the agent.
Audit Metadata