The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to read and process configuration files (package.json, netlify.toml) and command output (git remote show origin) to automate the deployment process. These files, which are often provided by the user or found in a project repository, could potentially be manipulated by an attacker to include malicious instructions that influence the agent's behavior.
Ingestion points: Project manifest files (package.json, netlify.toml) and Git remote information accessed during the deployment workflow in SKILL.md and references/deployment-patterns.md.
Boundary markers: None detected; the instructions direct the agent to analyze these files to detect frameworks and settings without specific delimiters or isolation.
Capability inventory: Shell command execution (npx netlify, npm install) and network communication with Netlify's infrastructure.
Sanitization: The skill does not specify procedures for sanitizing or validating the contents of the project files before processing them.
[EXTERNAL_DOWNLOADS]: The skill facilitates the download and execution of the Netlify CLI via npx netlify and project dependencies via standard package managers (npm, yarn, pnpm). These operations target the official NPM registry and Netlify's infrastructure, which are well-known services for this context.
[COMMAND_EXECUTION]: The instructions involve executing various shell commands to verify authentication (npx netlify status), link projects (npx netlify link), and perform deployments (npx netlify deploy).

netlify-deploy