market-news-analyst
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill utilizes "WebSearch" and "WebFetch" tools to ingest content from external financial news websites and official government sources. This creates a potential surface for indirect prompt injection, where malicious instructions embedded in a fetched webpage could influence the agent's behavior.
- Ingestion points: News collection workflow in "SKILL.md" (Step 1) utilizes web retrieval tools to gather data from the past 10 days.
- Boundary markers: Absent. The instructions do not include specific delimiters or directives to ignore instructions contained within the fetched news data.
- Capability inventory: The skill performs news retrieval and generates Markdown reports. It does not contain capabilities for shell execution, sensitive file access, or persistence.
- Sanitization: Absent. No specific sanitization or filtering of external content is defined in the workflow.
Audit Metadata