Add use-inlets-uplink skill

[welteki]

Description

Add the use-inlets-uplink skill to guide agents through managing Inlets Uplink tunnels using Kubernetes CRDs, the CLI, and REST APIs.

The skill covers:

• Tenant namespace and tunnel lifecycle management.
• Single and multiple TCP or HTTP upstreams.
• Mixed TCP and HTTP tunnels.
• Secure client connection and credential handling.
• Client, entitlement, and billing API guidance.
• Verification, troubleshooting, and cleanup.

Also adds skill interface metadata and lists the skill in the README.

Motivation and context

Provide agents with current guidance for creating, connecting, securing, and managing Inlets Uplink tunnels through its supported interfaces.

How has this been tested?

• Validated the skill structure with the skill validation tool.
• Checked the committed diff for whitespace and formatting errors.
• Tested against a running Inlets Uplink cluster using OpenCode with Qwen3.6-27B-FP8-vLLM.
• Exercised tunnel creation, CRD updates, client connection, resource verification, and cleanup.

[reviewfn]

AI Pull Request Overview

Disclaimer: This review was generated by automated AI and may contain errors. Do not trust its outputs without human verification.

Summary

• Adds a new use-inlets-uplink skill covering Kubernetes CRDs, the tunnel CLI plugin, REST APIs, billing, and safe credential handling.
• Adds interface metadata for OpenAI-compatible agents.
• Adds reference material for tunnel patterns and API/billing workflows.
• Updates the root README skill list with the new Uplink skill.
• The content is security-conscious and repeatedly avoids exposing tunnel tokens, licenses, and API keys.
• No blocking correctness, rendering, reproducibility, or publish-quality findings were identified in the scoped diff.

Approval rating (1-10)

9/10. The change is well-scoped, consistent with existing skills, and provides actionable Uplink guidance without obvious blocking defects.

Summary per file

Summary per file

File path	Summary
README.md	Adds use-inlets-uplink to the skills table.
skills/use-inlets-uplink/SKILL.md	Defines Uplink skill guidance for CRDs, CLI, API use, verification, and safe operations.
skills/use-inlets-uplink/agents/openai.yaml	Adds OpenAI interface display metadata and default prompt.
skills/use-inlets-uplink/references/api-and-billing.md	Documents client API endpoints, authentication handling, and billing API boundaries.
skills/use-inlets-uplink/references/tunnel-patterns.md	Provides TCP, HTTP, mixed tunnel, and token reference examples.

Overall Assessment

This PR is ready from a review perspective. The new skill follows the repository's established layout, the README entry is consistent with adjacent skills, and the reference documents give enough operational context for an agent to avoid common Uplink mistakes around namespaces, CRD qualification, generated connection output, and billing credentials. The examples are framed as templates with explicit placeholders, and the internal links resolve within the added skill directory.

Detailed Review

Detailed Review

Content review

No blocking findings.

The title, metadata, and README description align with the actual scope: the skill covers Kubernetes CRDs, CLI usage, REST APIs, and billing-related guidance rather than only tunnel creation.

The opening and structure are usable for an agent workflow: the main skill starts with namespace preparation and interface selection, then pushes detailed examples into references instead of overloading the top-level file.

The examples are reproducible as templates because placeholder values such as TENANT_NAMESPACE, UPLINK_DOMAIN, OBSERVED_SERVICE_ADDRESS, and OBSERVED_PORT are called out before use, reducing the chance that an agent applies them literally.

The credential-handling guidance is specific and actionable. It distinguishes product licenses, client API tokens, tunnel tokens, and billing API keys, and it repeatedly avoids printing captured connection output or secret values.

The API and billing reference makes the authorization boundary clear: billing capacity changes require explicit authorization and a separate billing API key, which is important because the adjacent tunnel-management API guidance could otherwise invite accidental endpoint or credential substitution.

AI agent details.

Agent processing time: 1m26.066s
Environment preparation time: 2.992s
Total time from webhook: 1m31.266s