trade-signal
Fail
Audited by Gen Agent Trust Hub on Mar 10, 2026
Risk Level: HIGHCOMMAND_EXECUTIONPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The
scripts/search.shscript is vulnerable to Python code injection. The$QUERYvariable is expanded directly into a Python command string using triple quotes ('''$QUERY'''). An attacker can use triple quotes in their input to break out of the string literal and execute arbitrary Python commands on the system. - [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface.
- Ingestion point:
scripts/search.shfetches data fromhttps://terminal-x.ai/api/lite-search. - Boundary markers: Absent.
- Capability inventory: The skill can execute local commands via
scripts/search.sh. - Sanitization: Absent. Malicious content in the API response could potentially exploit the agent's logic or the command injection bug.
- [EXTERNAL_DOWNLOADS]: The skill fetches financial data from
https://terminal-x.ai, which is the official homepage for the service.
Recommendations
- AI detected serious security threats
Audit Metadata