openai-netlify-deploy
Fail
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: HIGHREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION] (HIGH): The skill uses
npm installandnpx netlify, both of which execute code defined in the project'spackage.json(lifecycle scripts) ornetlify.toml(build commands). Malicious projects can use these to achieve RCE. - [COMMAND_EXECUTION] (HIGH): The skill uses
Bashto execute commands and explicitly suggests requesting escalated permissions for network access, which increases the potential impact of any malicious command execution. - [EXTERNAL_DOWNLOADS] (LOW): Fetches the Netlify CLI via
npx. While the download is from a trusted source (Netlify) per [TRUST-SCOPE-RULE], the subsequent execution on untrusted data maintains the high risk of the skill. - [PROMPT_INJECTION] (HIGH): High surface for Indirect Prompt Injection. 1. Ingestion points:
package.json,netlify.toml, and git remote URLs. 2. Boundary markers: Absent. 3. Capability inventory:Bashexecution,Writeaccess, and deployment script execution. 4. Sanitization: Absent; the skill blindly follows build and installation instructions provided in untrusted configuration files.
Recommendations
- AI detected serious security threats
Audit Metadata