Skip to main content

Agent Skills

Docs-hosted guidance for coding agents building with Basis Theory

Basis Theory Agent Skills are markdown instructions for coding agents. They help an agent choose the right Basis Theory primitive, ask the right implementation questions, and link back to the docs that define current product behavior.

The skills are committed in the Basis Theory developer docs and may change over time as APIs and integration patterns evolve. The public docs remain canonical. If a skill and a docs page disagree, use the docs page and treat the skill as stale.

Use the latest skills from the docs​

For most tasks, do not install anything. Point your agent at the hosted entrypoint so it reads the current pack each time:

Prompt
Read https://developers.basistheory.com/agent-skills/skills.md and use the relevant
Basis Theory Agent Skills while you help me build this integration.

/agent-skills/skills.md is the preferred starting point for agents. It lists every skill, the direct SKILL.md URL for each one, and the supporting markdown references. An agent can fetch that file, choose the skills that match your task, and then read only the relevant markdown files.

Agents that discover documentation automatically can also find the entrypoint through /llms.txt.

Optional: install the skills​

If your agent supports project-local skill files and you want the guidance available across repeated tasks, ask it to install the skills from this docs site:

Prompt
Install the Basis Theory Agent Skills from this docs site's /agent-skills/skills.md into this project.

The pack is plain markdown. Installing it should copy the entrypoint and linked SKILL.md / reference markdown files into the skills directory your agent already uses; it should not run code from the docs.

What is in the pack​

The pack currently includes a router, an implementation planner, and focused skills for common Basis Theory surfaces such as token collection, payments, Proxy, Reactors, 3DS, network tokens, account updater, wallets, issuing, PII, and agentic commerce.

  • bt-start — route a vague Basis Theory request.
  • bt-plan — plan architectural or PCI-scope-changing work.
  • bt-collect — collect sensitive data into tokens or token intents.
  • bt-payments — process payments through PSPs using tokens or Proxy.
  • bt-proxy — use Proxy for detokenization or inbound capture.
  • bt-architecture — choose tenants, applications, keys, permissions, and primitive boundaries.
  • bt-migrate — plan card and vault migrations with customer-owned proof gates.
  • bt-production — prepare production readiness.
  • bt-test-debug — debug by proving the failure boundary.
  • bt-3ds — implement 3DS and SCA flows.
  • bt-wallets — accept Apple Pay and Google Pay.
  • bt-card-lifecycle — keep stored cards chargeable over time.
  • bt-network-tokens — use network tokens in recurring programs.
  • bt-account-updater — handle account updater flows.
  • bt-reactors — use Reactors for secure programmable logic.
  • bt-pii — vault and govern PII and documents.
  • bt-issuing — display issued-card data safely.
  • bt-agentic — build Agentic Commerce flows.
  • bt-psp-orchestration — build customer-owned PSP adapter and portability plans.
  • bt-tokens — model token schemas and opaque references.

Feedback​

The skills are early and will keep improving. Contact us with what you were building and where the agent was wrong, thin, or outdated.

Last updated on