alpha-vantage

Pass

Audited by Gen Agent Trust Hub on Aug 8, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the user to install the requests and pandas libraries. These are official, well-known, and highly reputable packages from the Python Package Index (PyPI) used for data manipulation and HTTP communication.
  • [PROMPT_INJECTION]: The skill processes financial data retrieved from the Alpha Vantage API (alphavantage.co), which acts as an ingestion point for untrusted external content. This creates a surface for indirect prompt injection.
  • Ingestion points: External JSON data retrieved via the av_get function in SKILL.md.
  • Boundary markers: None identified; instructions do not advise the agent to treat API responses as untrusted data.
  • Capability inventory: The skill uses requests for network GET operations and print for outputting data. It does not possess high-risk capabilities like file system writes, shell command execution, or persistence mechanisms.
  • Sanitization: None; the skill processes the raw JSON response from the API.
  • [DATA_EXFILTRATION]: The skill requires the use of an ALPHAVANTAGE_API_KEY, which is transmitted to the official alphavantage.co domain for authentication. This is standard, intended functionality for an API wrapper and does not constitute unauthorized exfiltration.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 8, 2026, 12:54 PM
Security Audit — agent-trust-hub — alpha-vantage