Free, practical guides
Standalone technical writing on the same problems our kits solve — hooks, agent spend tracking, MCP servers. Each guide is genuinely useful on its own; the kits are for when you want the full implementation instead of building it from the concept.
Claude Code hooks: a practical guide
8 min readWhat hooks are, the PreToolUse/PostToolUse/Stop events, how to wire them in settings.json, a complete working command-logger example, and the pitfalls that quietly disable them.
Read guide →An append-only ledger for AI agent spend
7 min readWhy an autonomous agent's money trail needs to be tamper-evident by construction, the Postgres trigger pattern that enforces it, idempotency keys, and reconciliation.
Read guide →10 checks before you ship an MCP server
7 min readA distilled, genuinely useful pre-ship checklist: tool descriptions for LLM callers, actionable errors, untrusted-output injection, path traversal, rate limits, transport choice, versioning, and testing.
Read guide →