Auto-AI Web Bridge
PassAudited by VirusTotal on May 11, 2026.
Overview
Type: OpenClaw Skill Name: auto-ai-web Version: 1.0.0 The skill is classified as suspicious due to instructions in `SKILL.md` that involve high-risk operations. The agent is explicitly instructed to ask for and handle user credentials for external AI web services (Claude AI, Gemini), posing a risk of credential compromise. Additionally, the 'Comandi Utili' section provides a `curl` command demonstrating the agent's capability to download arbitrary files from any URL to the filesystem (`/home/rasberry/.openclaw/workspace/downloads/`), which could be exploited for arbitrary file write and potential remote code execution if combined with further malicious prompts.
Findings (0)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
A user could be prompted to give the agent passwords or account access for Claude/Gemini, potentially including a Google account, with unclear handling after use.
The skill instructs the agent to request third-party account credentials, use them for login, and potentially persist them, but it does not specify credential scope, storage location, retention, or safeguards.
Se NON loggato: chiedi credenziali all'utente ... Dopo aver ricevuto credenziali, prova il login ... Salva le credenziali in modo sicuro se l'utente le fornisce
Do not provide passwords to the agent. Prefer official APIs, OAuth/device login, or manual browser login, and require the skill to declare exactly what credentials are needed and how they are stored or discarded.
Private prompts, code snippets, or project details may be processed by Claude or Gemini through their web interfaces.
The skill sends user prompts, and potentially code or image-generation requests, to external provider web interfaces. This is purpose-aligned, but users should understand the data boundary.
Codice → Usa Claude AI (https://claude.ai) ... Immagini → Usa Gemini (https://gemini.google.com) ... inserisci il prompt
Only use this skill with data you are comfortable sending to those providers, and get explicit user approval before submitting confidential code or sensitive prompts.