analyst-common-stock

Pass

Audited by Gen Agent Trust Hub on Feb 22, 2026

Risk Level: SAFE
Full Analysis
  • SAFE (SAFE): The analyzed skill consists entirely of instructional rules and safety protocols intended to improve agent accuracy. It mandates the use of trusted financial sources (e.g., KRX, DART, Yahoo Finance) and prohibits the use of unverified data or simulated functions.
  • Indirect Prompt Injection (SAFE): The skill identifies a data ingestion surface where an agent reads user-provided markdown files from a local directory (materials_path). Under the adversarial reasoning framework, this is a standard functional surface mitigated by the skill's own strict verification rules.
  • Ingestion points: User-created .md files in the directory specified by the materials_path parameter.
  • Boundary markers: Absent in this common rules document; safety relies on the agent following the citation and verification protocols.
  • Capability inventory: The agent utilizes web search and fetching tools (exa_web_search_exa, websearch_web_search_exa, WebFetch) to verify data.
  • Sanitization: None explicitly defined, as the skill focus is on cross-referencing values rather than sanitizing file content.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 22, 2026, 05:46 PM