The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructions and the references/private-networking.md file direct the agent to fetch official platform documentation from https://render.com/docs/private-network.md and https://render.com/docs/private-network. These requests target the well-known service's official domain to ensure the agent uses the most current technical specifications.
[COMMAND_EXECUTION]: The skill recommends using tools such as curl or other HTTP clients to retrieve documentation. This execution is limited to fetching informational content from a trusted vendor for the purpose of assisting the user with platform configuration.
[INDIRECT_PROMPT_INJECTION]: The skill establishes an ingestion surface by instructing the agent to retrieve and read external Markdown and HTML files from the Render documentation site. This mechanism is designed to keep the agent's knowledge base current, and it relies on authoritative, well-known sources, which minimizes the risk of processing untrusted content.
Ingestion points: External URLs https://render.com/docs/private-network.md and https://render.com/docs/private-network mentioned in references/private-networking.md.
Boundary markers: None explicitly defined for the fetched content.
Capability inventory: The skill uses curl or HTTP clients to read documentation; it does not involve file system writes or subprocess execution for the fetched data.
Sanitization: The agent is instructed to read the content to confirm it is the expected document, acting as a manual validation step.

render-networking