aminer-deep-search
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill connects to the AMiner Open Platform API (https://datacenter.aminer.cn) to retrieve academic paper metadata and citation relations. This is a functional requirement for its stated purpose.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from an external API, creating a surface for indirect prompt injection.
- Ingestion points: Paper titles, abstracts, and author details are retrieved via
scripts/aminer_api.pyand passed to the agent. - Boundary markers: The
SKILL.mdfile provides clear instructions for the agent to act as a controller, judging relevance from structured JSON output. - Capability inventory: The skill uses
Bashand localpython3scripts to perform multi-round searches and state management. - Sanitization: Data is transferred as structured JSON; however, the natural language content within the fields is not sanitized for potential adversarial instructions.
Audit Metadata