TabTab Skill

Use TabTab to run AI-powered tasks in a sandboxed multi-agent environment. Supported capabilities: - General agent: open-ended tasks, writing, research, summ...

This skill appears to do what it says, but take these precautions before installing/using: 1) Verify the TabTab domain (default https://tabtabai.com) is the legitimate service you expect; if you set TABTAB_BASE_URL, point it only at a trusted instance. 2) Protect your API key: avoid committing scripts/env to source control; follow the script's chmod 600 suggestion, or use your OS secret manager rather than a plain file. 3) Review any files you upload (they are sent to the TabTab service). Don't upload sensitive secrets or private keys unless you trust the service. 4) Ensure required tools (curl, jq, stat) come from trusted packages on your system. 5) The scripts write temporary files under /tmp and expect a skill directory layout; adapt the .gitignore suggestion to your repo structure if needed. If you want additional assurance, inspect network traffic (or run in a restricted environment) to confirm the scripts only talk to the expected API endpoints.

Type: OpenClaw Skill Name: tabtab Version: 1.0.4 The 'tabtab' skill is a legitimate integration for the TabTab AI platform (tabtabai.com), providing a suite of shell scripts to manage remote multi-agent tasks. The skill includes scripts for authentication testing (hello.sh), file uploads (upload-files.sh), task creation (create-task.sh), and result retrieval (download.sh). It demonstrates good security practices, such as recommending restrictive file permissions (chmod 600) for secret storage, using arrays in curl commands to prevent shell injection, and performing client-side validation of file types and sizes before upload. No evidence of malicious intent, unauthorized data exfiltration, or prompt injection was found.