market-data
Pass
Audited by Gen Agent Trust Hub on Feb 18, 2026
Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
- Indirect Prompt Injection (LOW): The skill ingests untrusted news data (headlines and summaries) from the Finnhub API. This creates a surface where an attacker could influence the agent via poisoned news content.
- Evidence: 1. Ingestion points: News data fields in SKILL.md. 2. Boundary markers: Uses markdown templates but lacks specific instructions to disregard embedded commands in news strings. 3. Capability inventory: No executable scripts or system access. 4. Sanitization: No data validation or escaping mentioned.
- No Code (SAFE): The skill is entirely descriptive and instruction-based. It does not contain any executable scripts, shell commands, or external package dependencies.
Audit Metadata