market-sizing

Pass

Audited by Gen Agent Trust Hub on May 3, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection as it is instructed to ingest and process potentially untrusted external content.
  • Ingestion points: According to SKILL.md, the agent analyzes user-provided market research, industry reports, financial data, and results retrieved via web search.
  • Boundary markers: The instructions do not define explicit delimiters or include directives for the agent to ignore instructions embedded within the data being analyzed.
  • Capability inventory: The skill leverages web search tools to collect and analyze market data.
  • Sanitization: There are no defined procedures for sanitizing, validating, or filtering external content before it is incorporated into the agent's context.
Audit Metadata
Risk Level
SAFE
Analyzed
May 3, 2026, 02:37 PM
Security Audit — agent-trust-hub — market-sizing