The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill fetches documentation in markdown format from official sources.
Fetches documentation from https://neon.com/docs/introduction/branching.md and https://neon.com/docs/llms.txt to provide up-to-date context.
[REMOTE_CODE_EXECUTION]: The skill instructs the agent to install and execute official tools and extensions.
Executes npx -y neonctl@latest init --agent <agent-name> to set up the Neon CLI.
Executes npx -y add-mcp https://mcp.neon.tech/mcp to install the official Model Context Protocol (MCP) server.
Executes npx skills add neondatabase/agent-skills --skill neon-postgres-branches to install additional skill components from the official repository.
Installs the extension databricks.neon-local-connect for local development workflows.
[COMMAND_EXECUTION]: The skill utilizes shell commands for environment setup and package management.
Uses npm i and npm install for official packages such as @neondatabase/config and @neondatabase/env.
Interacts with local .env files to manage DATABASE_URL credentials, which is a standard and recommended practice.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection due to its reliance on external data.
Ingestion points: Reads markdown content from external URLs at neon.com and github.com (SKILL.md) to provide current documentation.
Boundary markers: The skill advises verifying claims against official docs but does not specify explicit delimiters for the fetched content within the agent's internal prompt context.
Capability inventory: The skill has capabilities to execute shell commands (npx), write to files (.env), and perform network operations (curl).
Sanitization: There is no explicit sanitization of the fetched markdown text before it is processed by the agent.

neon-postgres