personal-monitor-domain
Pass
Audited by Gen Agent Trust Hub on Jun 18, 2026
Risk Level: SAFENO_CODEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The domain specification describes a dashboard architecture that aggregates data from external, untrusted sources such as Gmail, Feishu, News feeds, and Social media. These inputs are subsequently processed by an LLM to generate 'AI Daily Briefing' summaries. This architecture presents a surface for indirect prompt injection, where malicious instructions hidden in emails or news articles could potentially influence the agent's summary generation.\n
- Ingestion points: Gmail API, Feishu API, GNews, HackerNews, RSS feeds, Twitter API, Xiaohongshu scraping.\n
- Boundary markers: The specification does not define any delimiters or safety instructions to isolate untrusted data from the LLM prompt.\n
- Capability inventory: The skill contains no executable code; the described dashboard utilizes network requests and handles sensitive authentication tokens.\n
- Sanitization: No sanitization or validation protocols are defined for external data before it is interpolated into LLM prompts.\n- [SAFE]: The skill consists entirely of documentation and domain specifications. It does not include any executable scripts, binary files, or installation commands that could be run on a host system.
Audit Metadata