Tooling

Use Skills from ~/.claude/skills/ when tasks match their purpose (e.g., /clojure-eval for evaluating Clojure code in the current REPL, /systematic-debugging for bug investigation, /go-testing for writing tests).
Use reef for codebase exploration — call reef query before reading files, reef impact before modifying functions, reef skeleton for file overviews. See /reef skill for full usage.
Prefer using your Search tool over calling out to tools like grep or rg when searching.
Use Mermaid diagrams to help explain complex systems and interactions.

Writing or editing Clojure

Before writing or editing any .clj/.cljs/.cljc file (including tests, schemas, HoneySQL queries, Integrant components, or REPL work), invoke the orcha-clojure-style skill. It covers naming, namespaces, formatting, testing patterns, Malli, type hints, database row builders, REPL reloading, and how to access running system components.

Git Commits

Before staging: Run git diff --name-only to see modified tracked files, and consider the task context for any new files you created that need to be added

Brainstorming specs

Read docs/superpowers/spec-design-system.html before writing one — it defines the tokens, components, and section order to use. Copy its <style> block verbatim.

Database Migrations

Important: Never edit resources/migrations/init.sql directly for schema changes. Always create new migrations to ensure changes are applied to existing databases.

AWS Profiles

orcha-prod: Production account (CDK stacks, Cognito, RDS, etc.)
orcha: Management account (Route53 delegation, cross-account resources)

Example: aws cognito-idp describe-user-pool --profile orcha-prod --region eu-central-1 --user-pool-id <id>

SES Email Receiving

For inbound email setup (DNS, receipt rules, debugging delivery issues), see docs/ses-email-receiving.md.

Database & S3 Access

PostgreSQL: psql -h localhost -U postgres -d orcha -c "..."
S3 (MiniStack): bb dev:aws-cli s3 ... (e.g., bb dev:aws-cli s3 ls s3://v1-orcha-global-storage-local-stack/)