Skip to content

Use Quant with AI coding agents

The Quant agent skills teach AI coding agents how to work with Quant: deploying static sites and Quant Cloud applications, writing rules and edge functions, setting up CI, purging caches, managing infrastructure as code, and using Search, AI and Studio. With the skills installed, an agent uses the right commands, endpoints and settings for your platform instead of guessing.

The repository holds 15 skills in the open Agent Skills format. They work with Claude Code, Codex and other agents that read skills or AGENTS.md. The skills are free to use under the MIT licence.

Run these two commands in Claude Code:

/plugin marketplace add quantcdn/agent-skills
/plugin install quant@quantcdn

The first command adds the Quant marketplace. The second installs the quant plugin, which contains all the skills. Claude Code loads a skill when your request needs it, so you do not have to name the skill.

To get the latest version later:

/plugin marketplace update quantcdn

Copy the skill folders from the repository into the folder your agent reads skills from:

Terminal window
git clone https://github.com/quantcdn/agent-skills.git
cp -r agent-skills/skills/* .claude/skills/ # or ~/.claude/skills, .agents/skills, ...

Agents that read AGENTS.md (for example Codex) can use the routing table in the repository’s AGENTS.md. Copy AGENTS.md and the skills/ folder into your project.

Quant runs as two platforms with the same products. The agent picks the right endpoints from the dashboard you sign in to:

You sign in to Platform
https://dashboard.quantcdn.io QuantCDN
https://dash.quantgov.cloud QuantGov

Say which one in your first request, for example “My project is on dash.quantgov.cloud”. If you do not say, the agent asks.

Most tasks need a token. Keep tokens out of the conversation:

  • Put tokens in environment variables or your CI secret store, and tell the agent the variable name, not the value.
  • Use the narrowest token that does the job: a project token for content deploys, and a dashboard API token for platform changes. Create API tokens in the dashboard under your profile: scope each token to one organisation, choose only the permissions the task needs, and set an expiry.
  • Review commands that change production (rules, domains, purges, deletes) before the agent runs them.

Start with requests like these:

  • “Deploy this Astro site to Quant with GitHub Actions.”
  • “Add a redirect rule from /old/* to /new/ on my Quant project.”
  • “Write an edge function that counts visits with a Quant KV store.”
  • “Move this Next.js app to Quant Cloud with a managed database.”
  • “Set up Terraform for my Quant project and import the existing rules.”
  • “Why does my proxy rule return a 502 Proxy Error?”

The quant-platform skill is the entry point. It explains the products and the two APIs, and routes to the other skills.

Skill Covers
quant-platform Products, QuantCDN and QuantGov, the content and platform APIs, which tool to use
quant-cli The quant content CLI: deploy, purge, edge functions, rules
qc-cli The qc cloud CLI: logs, environment variables, exec, backups
quant-static-sites Static sites: build output, deploys, unpublish, 404 pages, forms
quant-rules-and-routing Rules: matching, proxies, redirects, headers, WAF, bot challenge
quant-edge-functions-and-kv Edge functions, auth functions, content filters, KV stores
quant-cloud-apps Quant Cloud applications: compose, databases, variables, scaling, cron, backups
quant-templates Starter templates and writing your own
quant-ci-github-actions Quant GitHub Actions and other CI
quant-domains-cache-purge Custom domains, certificates, caching and purging
quant-iac Terraform and Pulumi
quant-cms-integrations The Drupal module and the WordPress plugin
quant-search Keyword search and AI Search
quant-ai Quant AI services: chat, tools, agents, workflows, governance
quant-studio Quant Studio, the visual editor for Astro sites

If a skill gives wrong or outdated advice, open an issue with the request you made and what the agent did.