deep-research
Pass
Audited by Gen Agent Trust Hub on Jul 16, 2026
Risk Level: SAFE
Full Analysis
- [EXTERNAL_DOWNLOADS]: Fetches scientific metadata and abstracts from the well-known OpenAlex API (
api.openalex.org). This is a legitimate research tool used to find academic papers. - [COMMAND_EXECUTION]: Uses standard shell commands (
mkdir,grep,curl) to manage the local research directory structure and fetch data from APIs. These operations are restricted to the research project scope. - [DATA_EXFILTRATION]: While the skill performs network operations and reads the local filesystem, the access is scoped to the
research/directory created by the skill and well-known research domains. No sensitive data exfiltration patterns were identified. - [PROMPT_INJECTION]: The skill uses the filesystem as 'working memory' to keep the context window clean. This architecture reduces the surface for direct prompt injection from large volumes of scraped text, although the agent should remain cautious when processing findings from untrusted web sources (indirect prompt injection surface).
Audit Metadata