companion-project-creator

Warn

Audited by Gen Agent Trust Hub on Feb 27, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands for project scaffolding and verification, including ecosystem-specific tools like composer, npm, pip, and go.
  • [REMOTE_CODE_EXECUTION]: The 'Verification Workflow' mandates the execution of generated code and unit tests on the host machine. This creates a high-risk surface for executing malicious code embedded in the source articles.
  • [EXTERNAL_DOWNLOADS]: The skill utilizes package managers to download and install external dependencies from public registries during the setup phase.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 27, 2026, 11:29 AM