exa-search
Pass
Audited by Gen Agent Trust Hub on Aug 17, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it fetches and processes untrusted data from the internet.
- Ingestion points: Untrusted content is retrieved through scripts/exa_search.py and scripts/exa_extract.py.
- Boundary markers: Instructions lack specific delimiters or warnings to ignore instructions embedded in the fetched data.
- Capability inventory: The skill allows file writing (via the -o flag) and network access (via the Exa API).
- Sanitization: No sanitization is performed on the retrieved content.
- [EXTERNAL_DOWNLOADS]: The skill dynamically downloads the exa-py package from PyPI using the uv run command.
- [COMMAND_EXECUTION]: The skill involves executing shell commands with interpolated user/agent arguments, which could be vulnerable to command injection if the inputs are not sanitized.
Audit Metadata