lov-publish-event-onto-hdx

Pass

Audited by Gen Agent Trust Hub on Sep 7, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external Huodongxing event pages, which is a potential attack vector for indirect prompt injection.
  • Ingestion points: The skill reads the ativityJson object and other DOM elements from live event pages on huodongxing.com via ego-browser (SKILL.md, Step 1).
  • Boundary markers: There are no explicit delimiters or specific instructions to the agent to disregard instructions potentially embedded within the ingested web data.
  • Capability inventory: The skill utilizes ego-browser for navigation and executes a local Python script (scripts/profile_store.py) to persist information to a local JSON profile.
  • Sanitization: While scripts/profile_store.py includes a blocklist to prevent saving sensitive keywords (e.g., 'token', 'secret', 'password') into the user profile, the skill does not implement specific sanitization to filter out malicious natural language instructions found in the processed web pages.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 7, 2026, 09:40 PM
Security Audit — agent-trust-hub — lov-publish-event-onto-hdx