gds-document-project

Fail

Audited by Gen Agent Trust Hub on Jul 7, 2026

Risk Level: HIGHCOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONCREDENTIALS_UNSAFE
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill executes a Python script located at {project-root}/_bmad/scripts/resolve_customization.py. Since this path is within the target project being analyzed, an attacker could include a malicious version of this script in a repository to gain code execution when a user runs the 'document project' command.
  • [COMMAND_EXECUTION]: The workflow performs deep-merging of configuration files, including {project-root}/_bmad/custom/{skill-name}.toml from the target project. It then executes commands listed in the activation_steps_prepend, activation_steps_append, and on_complete arrays sourced from these merged files, allowing the target project to dictate arbitrary shell commands for the agent to run.
  • [CREDENTIALS_UNSAFE]: As part of its standard documentation workflow, the skill is instructed to scan for and ingest the contents of .env files, configuration manifests, and files matching patterns like *auth*, *session*, and *jwt*. This systematically exposes sensitive credentials and security configurations to the agent's context.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Jul 7, 2026, 06:21 PM
Security Audit — agent-trust-hub — gds-document-project