gdelt-events-fetch
Pass
Audited by Gen Agent Trust Hub on Mar 15, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill downloads data from http://data.gdeltproject.org, which is the official endpoint for the GDELT Project. This is a well-known public data repository.
- [COMMAND_EXECUTION]: The skill executes a local Python script scripts/gdelt_events_fetch.py to manage the fetching process. The script is invoked with specific CLI arguments and uses environment variables for configuration.
- [DATA_EXFILTRATION]: The script writes downloaded data files and logs to the local filesystem (--output-dir, --log-file). This is the intended behavior for a data-fetching skill.
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests and displays external data.
- Ingestion points: scripts/gdelt_events_fetch.py reads data from lastupdate.txt, masterfilelist.txt, and the contents of downloaded ZIP files.
- Boundary markers: No delimiters or safety instructions are used when displaying the preview_lines output to the agent.
- Capability inventory: The skill can perform network requests and write to the local filesystem.
- Sanitization: While the script validates the structure of the ZIP files (CRC, column count, UTF-8), it does not sanitize the actual content of the records before outputting them, which could allow malicious instructions in the GDELT data to influence the agent.
Audit Metadata