ax-skill-creator

Create installable skill bundles with SKILL.md frontmatter and approved hosts.

1|Updated Apr 23, 2026
One-click install
npx skills add https://github.com/project-ax/ax-next --skill ax-skill-creator
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: ax-skill-creator
Source: https://github.com/project-ax/ax-next/tree/main/presets/k8s/src/builtin-skills/ax-skill-creator
Command: npx skills add https://github.com/project-ax/ax-next --skill ax-skill-creator

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It helps you create a safe, installable skill bundle for this assistant when you want to author an integration or workflow without quietly granting new capabilities.

Core Features & Use Cases

  • Guided skill authoring loop: Capture intent, draft a bundle under the required draft path, and iterate after testing.
  • Correct SKILL.md structure: Ensures the skill entrypoint includes YAML frontmatter with the required name and description, plus clear operational instructions.
  • Safety-aligned capability model: Keeps capabilities out of the frontmatter and routes hosts, credential slots, and packages through the approval step.
  • Validation rules awareness: Covers required grammar constraints like description length, skill id format, helper path rules, credential slot naming, and package/egress allowlisting.

Quick Start

Ask the assistant to create a skill by describing the capability you want, the external hosts and API key slot names it must use, and the public packages it would run, then follow the proposed authoring and install steps.

Frequently Asked Questions about ax-skill-creator

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I create an installable skill bundle for workflow automation securely?▼

To create an installable skill bundle securely, you define user intent and generate a draft package where external access to hosts, credential slots, and packages is routed through an approval step rather than granted directly.

What does YAML frontmatter require when authoring an integration skill?▼

YAML frontmatter requires only the name and description fields when authoring an integration skill, ensuring capabilities are kept out of the entrypoint definition and routed through security approvals.

How can I add API-backed capabilities to my assistant without granting unchecked access?▼

You can add API-backed capabilities by authoring a skill bundle that defines required external hosts, API key credential slots, and public packages, keeping them behind approved egress allowlisting and security approvals.

What validation rules apply to skill id format and credential slot naming during skill authoring?▼

Validation rules for skill authoring enforce specific grammar constraints including description length limits, skill id format patterns, helper path rules, credential slot naming conventions, and package or egress allowlisting requirements.

Can I include operational instructions directly in the skill entrypoint Markdown body?▼

Yes, you can and should include clear operational instructions directly in the Markdown body of the skill entrypoint, while keeping the YAML frontmatter limited to just the name and description fields.

Why does skill authoring require routing packages and hosts through an approval step?▼

Skill authoring requires routing packages and hosts through an approval step to maintain a safety-aligned capability model, preventing the quiet granting of new egress or integration capabilities without explicit authorization.