amazon-market-trend-scanner
Pass
Audited by Gen Agent Trust Hub on Aug 25, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a bundled Python utility
scripts/zoodata.pyto facilitate API interactions. The script relies entirely on standard libraries and includes a security check to ensure that the API key is only transmitted to trusted vendor domains (*.zoodata.ai) orlocalhost. - [DATA_EXPOSURE]: The skill manages state and credentials using local files such as
~/.zoodata/config.jsonand the{skill_base_dir}/scan-data/directory. This is documented behavior for persistent storage of user configuration and historical scan snapshots. - [INDIRECT_PROMPT_INJECTION]: The skill processes Amazon product reviews for sentiment and feature analysis, creating a potential surface for indirect prompt injection. The risk is mitigated through several best-practice safeguards.
- Ingestion points: Untrusted review text enters the agent context through the
reviews-rawandanalyzesubcommands withinscripts/zoodata.py. - Boundary markers: The skill uses triple-quote delimiters (
""") and explicit JSON schema instructions in its prompts to prevent the agent from interpreting review text as instructions. - Capability inventory: The skill's capabilities are limited to vendor API communication and local state file writes.
- Sanitization: The processing script automatically truncates review content to 500 characters, limiting the payload size for any potential injection.
Audit Metadata