data-scraper-agent
Pass
Audited by Gen Agent Trust Hub on Mar 31, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests data from external websites and processes it through an LLM. Evidence: 1. Ingestion points: Data enters the system via requests and playwright in the scraper/sources/ directory and main orchestrator. 2. Boundary markers: The prompt template in ai/pipeline.py uses Markdown headers for structure but does not include explicit instructions to ignore instructions found within the scraped data. 3. Capability inventory: The skill can perform network requests and write to Notion or other storage backends. 4. Sanitization: No sanitization or escaping is applied to the scraped content before it is included in the AI prompt.
- [EXTERNAL_DOWNLOADS]: The skill performs network operations to scrape user-defined websites and to interact with Google's Gemini API (generativelanguage.googleapis.com) for data enrichment. These operations are consistent with the skill's stated purpose.
Audit Metadata