gemini-workspace-bridge
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [Command Execution] (SAFE): The skill includes Bash snippets for creating directories (mkdir), initializing project files (cat), and performing file cleanup (find, rm). All commands are transparent, local, and appropriate for the stated purpose of workspace management.
- [Data Exposure & Exfiltration] (SAFE): The skill defines local directory structures for session artifacts. It does not access sensitive credentials, system-level configuration files, or perform any network operations.
- [Obfuscation] (SAFE): All instructions and scripts are provided in clear, human-readable text. No Base64 encoding, zero-width characters, or hidden payloads are present.
- [Prompt Injection] (SAFE): The content consists of architectural guidelines and technical instructions. There are no attempts to override system prompts or subvert AI safety constraints.
- [Indirect Prompt Injection] (SAFE): The skill establishes patterns for passing data between tools via local files (context bundles). While this represents a data ingestion surface, the implementation is local and uses standard file-based exchange methods without introducing specific vulnerabilities.
Audit Metadata