stock-evaluator-v3
Pass
Audited by Gen Agent Trust Hub on Mar 5, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: Comprehensive analysis of the 5 skill files did not reveal any malicious patterns, obfuscation, or security vulnerabilities.
- [DATA_EXPOSURE]: The skill retrieves investment data from well-known financial services and official regulatory domains. No sensitive file access, credential leakage, or unauthorized data exfiltration patterns were detected.
- [DYNAMIC_EXECUTION]: The skill generates a 'Quant-Style Dashboard' using a React artifact template. It uses the well-known 'recharts' library for data visualization. This dynamic generation is restricted to a predefined UI structure and does not involve the execution of untrusted remote scripts or arbitrary system code.
- [COMMAND_EXECUTION]: No dangerous system-level commands, privilege escalation (e.g., sudo), or persistence mechanisms were identified. The agent's activities are limited to standard web search and text/code generation within the safe environment.
- [PROMPT_INJECTION]: The instructions include robust defensive guardrails, such as a 'Zero Fabrication Policy' and strict data source hierarchies, which prevent the agent from hallucinating data or deviating from verified financial sources. No evidence of jailbreak attempts or instruction bypass was found.
Audit Metadata