gemini-antigravity-cli
Fail
Audited by Gen Agent Trust Hub on Jun 13, 2026
Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill provides instructions to download a binary archive from an external GitHub repository (testerlingcodo/gemini-antigravity-cli).
- [REMOTE_CODE_EXECUTION]: The installation guide instructs the user to unzip the downloaded file, apply execution permissions (chmod +x), and run the resulting binary, which involves executing code from an unverified external source.
- [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection as it is designed to ingest and process untrusted data from local files and directories using the /context command or shell piping.
- Ingestion points: File and directory content processed via the /context command and standard input piping in SKILL.md.
- Boundary markers: Absent. No specific delimiters or instructions to ignore embedded commands within the processed data are provided.
- Capability inventory: The skill can read local files, interact with the filesystem via MCP servers, and generate code based on context.
- Sanitization: Absent. The skill does not describe any validation or sanitization of the content ingested from files before it is processed by the AI model.
Recommendations
- AI detected serious security threats
Audit Metadata