All commands
Every command the published @codespar/cli exposes, by group, with the count each group carries.
@codespar/cli 0.16.0 exposes 134 commands, under 40 top-level entries.
This page enumerates. The task pages (Money, Mandates and governance, Catalog and discovery) teach, with a worked example and real output for each.
| Group | Commands | What it does |
|---|---|---|
| Direct commands | 20 | — |
| codespar servers | 2 | Browse the server catalog |
| codespar tools | 3 | Inspect tools exposed by servers |
| codespar consumers | 31 | Consumers: profile, Pix keys, Pix lookups, receipts, contact verification |
| codespar boletos | 4 | DDA boletos: subscribe a document, list the boletos it receives |
| codespar sellers | 6 | Sellers: onboarding status, custody, pending settlement, ledger |
| codespar mcp-servers | 8 | Tenant MCP servers: register, validate, list, patch a tool, sweep platform fees |
| codespar wallets | 15 | Wallets: balances, ledger, funding sources, execute, transfer, custody |
| codespar triggers | 12 | Webhook endpoints: deliveries, DLQ, secret rotation, redelivery |
| codespar charges | 5 | Charges: issue one, read it, list them, withdraw one |
| codespar consents | 4 | Consent tokens: mint the token whose URL the consumer opens to authorise an agent |
| codespar consumer-payments | 2 | Consumer payments: execute a payment on a consumer's behalf, with the audit chain (execute-stream returns SSE) |
| codespar catalog | 1 | Catalog: search it for a capability by intent |
| codespar payables | 5 | Payables: submit one (structured or manual), read its state and trail |
| codespar test | 4 | Sandbox money (test keys only): credit a consumer directly, or mint a Pix charge and settle it |
| codespar sessions | 3 | Inspect and manage sessions |
| codespar connect | 3 | Manage Connect Links for end-user provider OAuth |
| codespar mandate | 3 | Create and manage consumer mandates (the agent's allowance / wallet) |
| codespar audit | 1 | Read the tamper-evident audit chain |
| codespar logs | 1 | Inspect tool-call execution logs |
| codespar agent | 1 | Run an agent directory: reads its agent.yaml (schema 1) and delegates to the directory's own npm start |
What a command prints
Every command under a group page is derived from one HTTP operation and prints what that operation answered. The direct commands are written by hand and shape their own tables, but read --json the same way.
The route it called goes to stderr, so a pipe never sees it:
ℹ GET /v1/walletsOn stdout, the payload. A bare list, or an object with exactly one list field, becomes a table over the scalar fields its items share, up to six columns, and whatever else the payload carries prints as JSON below it:
id display_name currency status
wal_1 Example BRL active
{
"total": 1
}Everything else prints as JSON: an object with no list, one with two lists, or one whose single list came back empty. Nested fields never become columns. A value longer than the column prints clipped, and a line below the table says how many were clipped, so a cut cell is never silent.
--json replaces all of that with the payload verbatim, and is what a script should read. A 204 answer prints (no content) on stderr and nothing on stdout.
Exit codes and the shape of a failure are in Configuration.