skills/anbeime/skill/web-to-app/Gen Agent Trust Hub

web-to-app

Pass

Audited by Gen Agent Trust Hub on Jun 27, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes shell commands to verify system requirements and install the pake-cli tool globally using npm or pnpm. It also constructs and runs the pake build command with multiple parameters derived from user input in scripts/build_app.py.
  • [EXTERNAL_DOWNLOADS]: The skill downloads the pake-cli utility from the NPM registry during its initialization phase. This results in a global package installation on the host system.
  • [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface through its input parameters. Ingestion points: User-provided url, icon path, and inject file paths processed in scripts/build_app.py. Boundary markers: Absent; user inputs are interpolated directly into command-line arguments without delimiters. Capability inventory: Execution of shell commands via subprocess.run in scripts/build_app.py and scripts/install_pake.py. Sanitization: Absent; there is no validation or sanitization of file paths provided for icons or custom script injection, which could be abused to access or bundle sensitive local files into a redistributable application.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 27, 2026, 05:23 AM
Security Audit — agent-trust-hub — web-to-app