gstack
Fail
Audited by Gen Agent Trust Hub on Jun 28, 2026
Risk Level: HIGHCOMMAND_EXECUTIONCREDENTIALS_UNSAFEDATA_EXFILTRATIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill suite makes extensive use of bash scripts and binaries located in the ~/.claude/skills/gstack/bin/ directory for core operations, including preambles that run before AI interactions.
- [CREDENTIALS_UNSAFE]: The setup-browser-cookies functionality allows the agent to import active authentication sessions from the user's primary browsers, providing the AI with authenticated access to web accounts.
- [DATA_EXFILTRATION]: The pair-agent skill provides commands to initiate ngrok tunnels, exposing the local browser daemon to the public internet for remote agent coordination.
- [EXTERNAL_DOWNLOADS]: The browse tool setup can download and execute the bun runtime installer from https://bun.sh/install, although it implements SHA-256 checksum verification to ensure integrity.
- [REMOTE_CODE_EXECUTION]: The gstack-upgrade skill automates updates by pulling code from the author's GitHub repository and running a setup script.
- [DATA_EXFILTRATION]: The suite includes opt-in telemetry that sends anonymous usage statistics to a Supabase backend for project improvement.
Recommendations
- HIGH: Downloads and executes remote code from: https://bun.sh/install - DO NOT USE without thorough review
Audit Metadata