skills/vtmocanu/skills/skill-maker/Gen Agent Trust Hub

skill-maker

Pass

Audited by Gen Agent Trust Hub on Nov 1, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPERSISTENCEPRIVILEGE_ESCALATION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructions and scripts recommend using npx -y skills@latest to fetch the package manager and additional skills from remote repositories such as GitHub and GitLab.
  • [REMOTE_CODE_EXECUTION]: The refresh_skills.py script automatically downloads and installs code from external sources defined in lockfiles or command-line arguments. The skill documentation explicitly acknowledges that this allows third-party code to execute unattended.
  • [PERSISTENCE]: The skill provides detailed instructions to add SessionStart hooks to agent configuration files, specifically ~/.claude/settings.json and ~/.codex/hooks.json. This causes the update script to execute every time an agent session begins.
  • [COMMAND_EXECUTION]: The provided Python scripts (refresh_skills.py) use subprocess.Popen to execute shell commands for interacting with the npx CLI and managing git repositories.
  • [PRIVILEGE_ESCALATION]: The skill instructions guide the agent to modify its own configuration files and create directory symlinks in system and project-level directories, such as ~/.claude/skills and ~/.agents/skills, to ensure skill discovery across different agent platforms.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 1, 2026, 11:50 PM