agency-investment-researcher
Pass
Audited by Gen Agent Trust Hub on Jul 6, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is comprised entirely of persona definitions, instructional rules, and research templates. No security threats or malicious patterns were detected.\n- [PROMPT_INJECTION]: The instructions describe a workflow that includes processing external data from news, social sentiment, and web scraping. This represents a potential surface for indirect prompt injection, which is inherent to the skill's purpose as an investment researcher.\n
- Ingestion points: SKILL.md identifies earnings transcripts, news sources, social sentiment, and alternative web data as ingestion points.\n
- Boundary markers: The instructions do not include specific delimiters or boundary markers to isolate untrusted external content.\n
- Capability inventory: The skill mentions technical capabilities such as Python data analysis (pandas, yfinance) and web scraping, but no executable code or scripts are included.\n
- Sanitization: There are no explicit instructions for the sanitization or validation of external data before analysis.
Audit Metadata