The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes project source code, build scripts, and configuration files (such as netlify.toml and .env.production) during the deployment process, which presents a surface for indirect prompt injection.
Ingestion points: Project source files, build configuration files, and environment variable files within the SKILL.md instructions.
Boundary markers: No explicit delimiters or warnings regarding untrusted embedded instructions are defined for the content being deployed.
Capability inventory: The skill executes shell commands for deployment (netlify deploy), manages environment variables (netlify env:set), and writes status updates to progress.txt.
Sanitization: The skill relies on standard CLI tools and does not perform independent sanitization of project-provided content before processing.
[SAFE]: The skill uses the official netlify-cli and official plugins (like @netlify/plugin-nextjs) from Netlify, a well-known and established service provider. Downloads and installations from this source are standard and expected.
[SAFE]: Credential handling follows security best practices by utilizing environment variables (NETLIFY_AUTH_TOKEN) and importing secrets from .env files rather than hardcoding sensitive data into the skill or scripts.

vibe-coding-deploy-netlify