skills/wulaosiji/skills/rss-feed/Gen Agent Trust Hub

rss-feed

Pass

Audited by Gen Agent Trust Hub on Apr 18, 2026

Risk Level: SAFEDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [DATA_EXFILTRATION]: The script rss_fetcher.py contains a hardcoded absolute file path /Users/delta/.openclaw/workspace/01-Projects/daily-report/01-raw-materials/rss-data. This exposure reveals the local system's username ('delta') and internal project directory structure to anyone with access to the skill code.\n- [PROMPT_INJECTION]: The skill processes untrusted content from multiple external RSS feeds, which presents an indirect prompt injection surface.\n
  • Ingestion points: External data is fetched from various URLs (e.g., Google News, TechCrunch) in rss_fetcher.py.\n
  • Boundary markers: The resulting JSON output lacks any boundary markers or instructions to downstream agents to ignore embedded commands.\n
  • Capability inventory: The script performs network reads and local file writes via the RSSFetcher class.\n
  • Sanitization: There is no sanitization or filtering of the article summaries or titles before they are written to disk, allowing potential malicious instructions from a compromised RSS source to be stored.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 18, 2026, 01:11 AM
Security Audit — agent-trust-hub — rss-feed