web-deploy
Pass
Audited by Gen Agent Trust Hub on Mar 7, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references the use of CLI tools and packages from well-known services including Cloudflare (wrangler), Vercel, Netlify, and Docker. These are industry-standard tools from trusted organizations.
- [COMMAND_EXECUTION]: Provides instructional examples for executing build and deployment commands such as 'docker build', 'npx vercel', and 'npx wrangler'. These commands are consistent with the skill's primary purpose of automating web deployments.
- [CREDENTIALS_UNSAFE]: The documentation correctly identifies the need for authentication tokens (e.g., $CLOUDFLARE_API_TOKEN, $VERCEL_TOKEN) and demonstrates using environment variables and GitHub Secrets rather than hardcoding sensitive information.
- [INDIRECT_PROMPT_INJECTION]: As the skill involves processing project configuration files (e.g., vercel.json, netlify.toml) and repository content, it possesses an inherent attack surface where malicious data in those files could influence the deployment process.
- Ingestion points: Project configuration files (vercel.json, .github/workflows/deploy.yml) and source code in the project directory.
- Boundary markers: None specified in the instructional examples.
- Capability inventory: Execution of shell commands via CLI tools and network operations via API calls (fetch).
- Sanitization: No explicit sanitization of input files is described in the provided examples.
Audit Metadata