The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to run npx neonctl@latest init to set up the environment, install VSCode extensions, and configure MCP servers.
Evidence: Found in references/devtools.md and references/getting-started.md. This is a standard initialization command for the vendor's official CLI tool.
[EXTERNAL_DOWNLOADS]: The skill fetches documentation and index files directly from the vendor's official domain.
Evidence: SKILL.md suggests fetching markdown from https://neon.com/docs/introduction/branching.md and the docs index from https://neon.com/docs/llms.txt. These are trusted vendor resources.
[DATA_EXPOSURE]: The skill provides instructions for reading and modifying local .env files to store DATABASE_URL and NEON_AUTH_COOKIE_SECRET.
Evidence: Found in references/getting-started.md. The skill includes safety reminders to use environment variables and not commit connection strings to version control.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface where it processes data from the user's local codebase and environment to determine the integration status.
Ingestion points: Inspects existing connection code, ORM configurations, and .env files (references/getting-started.md).
Boundary markers: No specific delimiters or boundary warnings are mentioned for the codebase inspection.
Capability inventory: The skill can execute SQL queries, manage database branches, and provision auth resources via the Neon CLI or MCP server (references/neon-cli.md, references/devtools.md).
Sanitization: Not explicitly specified for the codebase analysis phase.

neon-postgres