trailmark-structural

Warn

Audited by Gen Agent Trust Hub on Apr 29, 2026

Risk Level: MEDIUMCOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The shell command in Step 2 (find {args} -type f ...) interpolates the user-provided {args} variable without shell-safe quoting. This allows an attacker to execute arbitrary shell commands by supplying a path containing metacharacters (e.g., ; rm -rf /).\n- [COMMAND_EXECUTION]: The shell command in Step 3 (trailmark analyze ... {language_flag} {args}) also interpolates {args} without quoting. This provides an additional vector for command injection, allowing arbitrary code execution with the permissions of the agent process.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Apr 29, 2026, 09:07 PM
Security Audit — agent-trust-hub — trailmark-structural