stripe-apilisted

# Stripe API Every Stripe endpoint the pinned stripe-node SDK exposes is reachable through the bundled engine. Do not hand-write HTTP calls. ## Command grammar ``` ${CLAUDE_PLUGIN_ROOT}/bin/stripe-x <resource.path> <action> \ [--account <name|comma-list|all>] [--live] \ [--id <id>] [--params '<json>'] [--data k=v --data a[b]=c] \ [--expand a,b] [--limit N | --all] \ [--idempotency-key K] [--api-version YYYY-MM-DD] \ [--confirm] [--confirm-bulk] [--arm-live] [--table] ``` `resource.path` is the dotted SDK path in snake or camel case: `customers`, `checkout.sessions`, `issuing.cards`, `treasury.financial_accounts`, `reporting.report_runs`. `action` is `list`, `retrieve`, `create`, `update`, `del`, `search`, `cancel`, `capture`, and so on. ## Rules 1. Reads (`list`, `retrieve`, `search`) run immediately and may target `--account all` or a comma list for a consolidated, per-account-tagged result. 2. Writes need a single `--account`. The engine prints a CONFIRMATION REQUIRED preview and exits without calling Stripe unless `--confirm` is passed. Relay the preview to the user and only re-run with `--confirm` after they approve. 3. Test mode is default. Live mode needs `--live`. Arming is a separate call: `--live --arm-live` arms the session for that account and exits 14 without executing. Only then does `--live --confirm` run the operation, and only with explicit approval in the same turn. 4. Bulk writes use `--bulk-ids id1,id2,...`. Over the threshold the engine