Update from code changes: runpodctl agent-friendly output

[mintlify]

Summary

Document the new agent-friendly output introduced in runpodctl v2: GPU pricing plus per-data-center availability on gpu list, ready-to-call invoke URLs on serverless create/get/list, and the CLI-wide JSON error contract with stable code values.

Changes

• runpodctl/reference/runpodctl-gpu.mdx: Updated the gpu list example to include securePricePerHr and communityPricePerHr (nullable when a cloud type does not offer the GPU) and the new dataCenterAvailability array. Added an output-fields section explaining stockStatus (best overall), available, and the "offered here, out of stock" vs. "not offered here" distinction.
• runpodctl/reference/runpodctl-serverless.mdx: Added an "Invoke URLs in command output" section showing the new urls object (run, runsync, health) emitted by serverless create/get/list, noting the URLs are computed client-side and that RUNPOD_INVOKE_URL targets a non-production invoke host.
• runpodctl/overview.mdx: Added a new "Output format" section covering the JSON-by-default contract, the [] empty-result behavior for template search and hub search, the flat JSON error shape on stderr with the stable code vocabulary (usage_error, not_found, no_credentials, network_error, cli_error, and REST/GraphQL-derived codes), the exit-code behavior change for model, update, send, and receive, and an environment variables table including the new RUNPOD_INVOKE_URL.

Context

Follows the upstream runpodctl PR that introduced the agent-friendly output shape (CON-683, CON-684, CON-685).

[mintlify]

Preview deployment for your docs. Learn more about Mintlify Previews.

Project	Status	Preview	Updated (UTC)
runpod-docs	🟢 Ready	View Preview	Aug 4, 2026, 10:12 PM