db-reset

Pass

Audited by Gen Agent Trust Hub on Feb 19, 2026

Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
  • [Prompt Injection] (LOW): The skill is susceptible to indirect prompt injection because it instructs the agent to identify and execute commands derived from 'project tooling'.
  • Ingestion points: Local project configuration files and tooling output.
  • Boundary markers: None mentioned in the workflow to separate data from instructions.
  • Capability inventory: Execution of destructive shell commands (reset/seed).
  • Sanitization: No validation or sanitization of identified command patterns is described.
  • [No Code] (SAFE): No scripts, binaries, or executable logic files were found in the skill.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 19, 2026, 01:32 AM