plan-fix-linter-error
Fail
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: HIGHCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION] (HIGH): The skill performs shell command construction using unvalidated user input. Evidence: In Step 2 (Verification), the command
bun run lint 2>&1 | grep -E "($ARGUMENTS)" | wc -ldirectly interpolates$ARGUMENTS. Risk: An attacker providing input likerule-name); malicious_command; #could achieve arbitrary code execution on the host system. - [PROMPT_INJECTION] (MEDIUM): The skill lacks input validation or sanitization for the
$ARGUMENTSvariable, making it susceptible to indirect prompt injection where data processed (the rules) can manipulate the agent's shell execution. Ingestion points:$ARGUMENTSvariable. Boundary markers: Absent. Capability inventory: Bash execution. Sanitization: Absent.
Recommendations
- AI detected serious security threats
Audit Metadata