startup-it-troubleshooting
Fail
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: HIGHREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONCREDENTIALS_UNSAFEEXTERNAL_DOWNLOADSPRIVILEGE_ESCALATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill contains patterns that download and execute scripts directly from remote sources.
SKILL.md: Executes a script fromgam-shortn.appspot.comusing bash process substitution:bash <(curl -s -S -L https://gam-shortn.appspot.com/gam-install).SKILL.md: Executes the Homebrew installation script using command substitution:bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)".- [CREDENTIALS_UNSAFE]: The skill includes hardcoded temporary passwords in example commands which could lead to security risks if deployed in production.
SKILL.md: Hardcoded passwordTempPass123!in a Google Workspace user update command.SKILL.md: Hardcoded passwordWelcome2Company!in a user creation command.- [PRIVILEGE_ESCALATION]: Several commands attempt to elevate privileges or enable remote administrative access.
SKILL.md: Enables remote SSH login usingsudo systemsetup -setremotelogin on.SKILL.md: Enables Remote Desktop (RDP) on Windows by modifying the registry (fDenyTSConnections).SKILL.md: Multiple uses ofsudoto modify network configurations, flush DNS caches, and manage system services.- [COMMAND_EXECUTION]: The skill makes extensive use of powerful shell and PowerShell commands to manage users, account security, and system states.
SKILL.md: Commands for modifying user MFA, backup codes, and account suspension status in identity providers.SKILL.md: Commands for terminating processes (pkill,Stop-Process) and managing storage via Docker.- [EXTERNAL_DOWNLOADS]: The skill installs external software and modules without version pinning or integrity checks.
SKILL.md:pip install speedtest-cli.SKILL.md:Install-Modulefor Microsoft 365 and Windows Update management.- [INDIRECT_PROMPT_INJECTION]: The "First-Day Setup Script" creates an ingestion point for user-supplied data used in shell commands.
- Ingestion point: User input via
readcommands for name and email inSKILL.md. - Boundary markers: None.
- Capability inventory:
git,gh,sudo,bash. - Sanitization: None.
Recommendations
- HIGH: Downloads and executes remote code from: https://gam-shortn.appspot.com/gam-install, https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh - DO NOT USE without thorough review
- AI detected serious security threats
Audit Metadata