tzai-banner

Pass

Audited by Gen Agent Trust Hub on Aug 10, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses a shell script to locate an executable engine within various hidden directories in the user's home folder ($HOME/.agents, $HOME/.claude, etc.) and executes it using bash "$ENGINE". While these paths are targeted at related vendor skills, executing scripts from computed paths in user directories carries a risk of executing unauthorized code if those paths are compromised.
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection and command injection. User-supplied text (referred to as <user subject>) is interpolated directly into a shell command: bash "$ENGINE" banner --prompt "<user subject>".
  • Ingestion points: User-provided subject/prompt text in SKILL.md.
  • Boundary markers: None. The input is placed directly inside double quotes within a bash command.
  • Capability inventory: The skill has the capability to execute shell commands via bash.
  • Sanitization: None detected. An attacker could provide a payload like \"; touch /tmp/pwned; # to execute arbitrary commands.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 10, 2026, 02:49 AM
Security Audit — agent-trust-hub — tzai-banner