adbertram

mercari-cli

Use this skill for service operations only. DO NOT use this skill for CLI implementation lifecycle work such as creating, testing, updating, troubleshooting, validating, removing, or documenting the CLI tool itself; delegate those tasks to cli-tool-expert. MANDATORY: Execute Mercari operations using the `mercari` CLI tool. CLI interface for Mercari. Triggers: mercari, mercari cli

adbertram 6 Updated 1mo ago

Resources

1
GitHub

Install

npx skillscat add adbertram/cli-tools/mercari-cli

Install via the SkillsCat registry.

SKILL.md
Execute Mercari operations using the `mercari` CLI. All Mercari interactions should use this CLI. The `mercari` CLI follows this pattern: ```bash mercari [arguments] [options] ```
Task Command
Authenticate mercari auth login
Check auth mercari auth status
Search others' listings mercari listings search lego --table --limit 10
Search sold, cheapest first mercari listings search lego --status sold --sort price --table
Search newest listed first (default) mercari listings search lego --sort newest --limit 20 --table
Filter search by condition/price mercari listings search lego --condition good --min-price 20 --max-price 100 --table
List active listings mercari listings list --status active
List sold listings (table) mercari listings list --status complete --table
List inactive listings mercari listings list --status inactive --limit 25
Get one listing/item mercari listings get m12345678901
Get multiple listings/items mercari listings get-many m12345678901 m10987654321
</quick_start>
**MANDATORY: Verify the live command shape before executing ANY `mercari` command.** Consult `usage.json` when the repo or installed package ships it. If `usage.json` is absent, use `mercari --help`, the relevant subcommand `--help`, and `README.md` instead. Never guess at command syntax. - **listings** -- Read and search Mercari listings (read-only): - `search ` -- OTHER sellers' public listings. Options (all validated against the real search API): `--status on_sale|sold` (long-form only here), `--condition new|like_new|good|fair|poor`, `--min-price`/`--max-price` (US dollars), `--sort/-s newest|price|relevance` (default `newest`; Source-CLI Sort Standard) with `--desc/-d` to reverse (only valid with `price` — Mercari has no oldest-first, so `newest --desc`/`relevance --desc` error), `--category-id`/`--brand-id` (repeatable), plus `--limit`, `--filter`, `--properties`, `--table`. Each result has `id` + canonical `url`. - `list` -- the authenticated seller's OWN listings (by `--status active|inactive|complete`, with `--limit`, `--filter`, `--properties`, `--table`). - `get ` -- full detail for any listing/item (by id or URL, with `--properties`, `--table`). It reports `buyer_protection_fee_cents` from Mercari's explicit `priceSummary.headline` and `landed_total_cents` from numeric `priceSummary.totalPrice`. The landed total can include buyer-paid shipping and other checkout costs. Both fields use integer cents. - `get-many ` -- full detail for up to 25 items through one browser session. It preserves input order and coverage. Each result uses `status: ok` with `item`, or `status: error` with `error_kind` and `error`. Oversized batches fail before opening the browser. Use `--status-only` for uncached availability fields (`id`, `url`, `status`, and `lastSoldAt` when present) without requiring checkout pricing. Full-detail pricing remains strict, and a pricing parse failure returns a keyed `unreadable` item error while preserving valid siblings. A human verification challenge stops the command. - **auth** -- Authentication management (login, logout, status, test) and nested `auth profiles` - **cache** -- Local response cache management This CLI is read-only. It drives the authenticated Mercari web app and captures its internal GraphQL responses; there are no mutating commands. Login requires an email one-time verification code. </essential_principles> **`usage.json`** -- Complete command tree with arguments, options, defaults, and usage instructions when present. **`mercari --help` and subcommand `--help`** -- Live installed command tree and option list. **`README.md`** -- Supplemental examples and workflow notes. </reference_index> - Command executes without error - Output is displayed in requested format - Correct command and flags used, verified against the live help output or `usage.json` when present </success_criteria>

Categories