oro-behat-debugging
Warn
Audited by Gen Agent Trust Hub on Aug 4, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides instructions to execute high-privilege shell commands using
sudoto modify system configuration files, including/etc/fstaband PostgreSQL configuration directories, for performance tuning (found inreferences/performance-tmpfs.md). - [DATA_EXFILTRATION]: The 'breadth-first diagnostics' pattern in
references/breadth-first-diagnostics.mdinstructs the user to create diagnostic steps that dump extensive database state (including visibility tables, price lists, and checkout workflows) into the console output, which may lead to sensitive data exposure if these logs are captured by shared CI/CD systems or logging aggregators. - [PROMPT_INJECTION]: The skill contains an indirect prompt injection surface by instructing the agent to ingest and analyze potentially untrusted data sources including application server logs (
var/log/<env>.log) and browser console logs (references/ajax-flake.md). - Ingestion points: Application logs (
var/log/<env>.log), browser console logs. - Boundary markers: None mentioned.
- Capability inventory: Execution of Behat CLI commands, database querying (via provided PHP code stubs), and browser session interaction.
- Sanitization: None described for external log content.
Audit Metadata