semantic-scholar-database
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements a legitimate wrapper for the Semantic Scholar Graph API. Analysis of the scripts and instructions revealed no malicious intent, obfuscation, or unauthorized access patterns.
- [EXTERNAL_DOWNLOADS]: The skill performs network requests to
api.semanticscholar.org. This is a well-known scholarly service, and the requests are limited to the official API endpoints for searching and retrieving paper metadata. - [CREDENTIALS_UNSAFE]: API authentication is handled securely via the
S2_API_KEYenvironment variable. The skill correctly usesos.environ.get()and does not contain hardcoded secrets or sensitive credentials. - [INDIRECT_PROMPT_INJECTION]: The skill processes external data (paper abstracts, titles, and author details) retrieved from the Semantic Scholar API. While this creates an ingestion surface for potentially untrusted content, the risk is minimal as the skill only performs data retrieval and does not interpolate this data into commands or high-privilege operations.
- [METADATA_POISONING]: The inclusion of an audit result file (
semantic-scholar-database_audit_result_v1.json) contains self-referential safety claims. While these claims were evaluated as data rather than authoritative conclusions, the content of the file does not contain deceptive instructions or injection patterns.
Audit Metadata