Skills
Navigate
skills/calionauta/agent-sync-public/stelow-product-orchestrator/Gen Agent Trust Hub

stelow-product-orchestrator

Warn

Audited by Gen Agent Trust Hub on Jul 18, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill provides instructions for the agent to install and execute code from various external, unverified repositories and packages at runtime.
  • Reference in references/cli-tools/agent_browser.md instructs running npx -y @earendil-works/pi-agent-browser, which downloads and executes the package immediately.
  • Reference in references/cli-tools/codequality-review.md instructs the agent to install external plugins using pi install git:github.com/cursor/plugins and npx skills add cursor/plugins -g.
  • Reference in references/cli-tools/safe-change.md suggests installing pi install git:github.com/PriNova/pi-agent-codebase-workflows.
  • Reference in references/cli-tools/todo.md instructs installation of @tintinweb/pi-tasks and @juicesharp/rpiv-todo from NPM.
  • Reference in references/cli-tools/subagents.md mentions @tintinweb/pi-subagents and pi-subagents.
  • [COMMAND_EXECUTION]: The skill uses extensive shell commands and node -e scripts to manage workflow state, file operations, and tool coordination.
  • SKILL.md and stages/setup.md use complex node -e scripts to read and write to stelow.json.
  • Frequent use of bash, cat, grep, and git commands throughout the stage files (e.g., stages/execution.md, stages/gate.md) to manipulate the filesystem and codebase.
  • [EXTERNAL_DOWNLOADS]: The skill facilitates downloading tools and configurations from non-whitelisted GitHub organizations including cursor, PriNova, juicesharp, and tintinweb.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jul 18, 2026, 10:48 AM
Security Audit — agent-trust-hub — stelow-product-orchestrator