Each playbook answers one question end to end, using the MCP tools or the REST API,
and shows a real executed transcript rather than an idealised one — including where
the first answer needed verifying.

* [Find a subnet for my app](/docs/playbooks/find-a-subnet-for-my-app) — a
  plain-language task to candidate subnets, then to calling code.
* [Evaluate a subnet before staking](/docs/playbooks/evaluate-a-subnet-before-staking)
  — what to check, and what the numbers do and do not tell you.
* [Monitor my validator](/docs/playbooks/monitor-my-validator) — the surfaces that
  say whether a validator is healthy and earning.
* [Audit an account history](/docs/playbooks/audit-an-account-history) — reconstruct
  what one SS58 address has done on-chain.