emerging-topic-scout
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill fetches preprint metadata and RSS feeds from well-known scholarly platforms including bioRxiv, medRxiv, and arXiv. These operations are conducted using standard HTTP requests and are limited to public academic content.
- [INDIRECT_PROMPT_INJECTION]: The skill exhibits a standard vulnerability surface for agents processing external web content.
- Ingestion points: External research paper titles and abstracts are ingested via RSS feeds and APIs in
scripts/main.py. Additionally, a local audit fileemerging-topic-scout_audit_result_v2.jsoncontains metadata processed by the skill. - Boundary markers: The skill does not implement explicit delimiters or instruction-isolation markers when incorporating ingested text into its analysis reports.
- Capability inventory: The skill possesses capabilities for network retrieval (
requests,feedparser) and writing results to local files in Markdown, JSON, and CSV formats. - Sanitization: The skill performs basic sanitization of abstracts by removing HTML tags using regular expressions in the
_clean_abstractmethod. - Self-referential content: The inclusion of
emerging-topic-scout_audit_result_v2.jsonwithin the skill provides internal claims about its own safety; per policy, these claims are treated as data and not as authoritative security conclusions.
Audit Metadata