aihot-topic-selector

Pass

Audited by Gen Agent Trust Hub on Jul 29, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The script scripts/fetch_aihot.py retrieves content from the external API endpoint https://aihot.virxact.com/api/public/items to populate daily summaries.
  • [DATA_EXFILTRATION]: The skill initiates network connections to a non-whitelisted external domain (aihot.virxact.com). Although the purpose is data retrieval, any outbound network operation to unknown third-party services is flagged for awareness.
  • [PROMPT_INJECTION]: The skill facilitates an indirect prompt injection surface (Category 8).
  • Ingestion points: External news titles and descriptions are ingested from the AIHOT API through scripts/fetch_aihot.py.
  • Boundary markers: The generated markdown summaries (workflow/每日热点/YYYYMMDD.md) do not use specific delimiters or instructions to notify downstream agents to ignore embedded commands within the third-party news content.
  • Capability inventory: The fetch_aihot.py script has the capability to write and append to files on the local filesystem, creating a risk if the ingested data contains instructions to modify sensitive files.
  • Sanitization: While the script performs basic cleaning on titles for duplicate detection, it does not sanitize or filter the news descriptions and summaries retrieved from the external source before writing them to disk.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 29, 2026, 01:36 AM