Traversing Citation Networks
Pass
Audited by Gen Agent Trust Hub on Oct 28, 2026
Risk Level: SAFE
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill fetches paper metadata, references, and citations from the Semantic Scholar API (api.semanticscholar.org), which is a well-known and trusted service for academic research data.
- [COMMAND_EXECUTION]: Uses standard curl commands to perform HTTP GET requests to the Semantic Scholar API. These commands are limited to data retrieval and do not execute remote code.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from an external API, creating a surface for indirect prompt injection.
- Ingestion points: Data such as paper titles, abstracts, and citation contexts are ingested from api.semanticscholar.org responses.
- Boundary markers: No explicit delimiters or boundary markers are used to isolate external content within the agent's context.
- Capability inventory: The skill performs network operations via curl and writes metadata to local JSON files (papers-reviewed.json, citation-graph.json).
- Sanitization: Uses a scoring system based on keyword matches in titles, abstracts, and contexts to filter out irrelevant or potentially malicious content before it is processed further.
Audit Metadata