trade-analyze

Pass

Audited by Gen Agent Trust Hub on May 16, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill operates as a legitimate financial analysis orchestrator, using internal tools to collect and synthesize public market data.
  • [DATA_EXPOSURE_AND_EXFILTRATION]: No sensitive local files, credentials, or environment variables are accessed. The skill only processes public financial data and writes a summary report to the local directory.
  • [PROMPT_INJECTION]: The instructions provide a clear framework for the AI agent to act as a financial analyst. There are no attempts to bypass safety filters or override core instructions.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted web data.
  • Ingestion points: WebSearch results in SKILL.md.
  • Boundary markers: Absent; data is organized into a DISCOVERY_BRIEF but without isolation warnings.
  • Capability inventory: Writes report files to the filesystem and launches sub-agents.
  • Sanitization: Absent. While the surface exists, the risk is localized to the generated report content.
Audit Metadata
Risk Level
SAFE
Analyzed
May 16, 2026, 01:20 PM
Security Audit — agent-trust-hub — trade-analyze