AssetClaw Asset Management Assistant

Problem

Asset management systems usually split authentication, tenant context, query fields, approval states, and write-operation safety across different APIs. When engineers or operations staff hand-write curl calls, failures often come from expired Tokens, ambiguous tenant selection, missing Idempotency-Key, Chinese query encoding, or unconfirmed asset codes. AssetClaw packages common asset-ops workflows around these AssetHub HTTP APIs.

How It Works

The skill talks to AssetHub over HTTP and follows a login-first model: at the start of a session it calls the login endpoint, stores the Bearer Token and tenant context in a session file, and clears the cached credentials when logout is requested. For multi-tenant accounts, it lists the available enterprises or tenants and lets the user pick one by index, or locks the context to an externally provided tenant_id.

The main workflows include:

• Querying: use modules, module, and request GET to discover endpoints and read data, with results suitable for table-formatted output.
• Writing: check before writing and re-query after writing, attach Idempotency-Key to write requests, handle second-confirmation flows for high-risk endpoints, and use the AI safety entry point where supported.
• Coverage: assets, maintenance, inventory, transfer, idle assets, scrapping, procurement, quality control, acceptance, documents, depreciation, departments, users, roles and permissions, IoT, asset location, labels, alerts, audit logs, dashboards, and workflows.

Limits

The skill depends on the backend API being reachable, the user having the right tenant and role permissions, and live API responses taking priority over static docs. Chinese search parameters need URL encoding. Department-level assets may be distributed across department, location, and use_department fields, so single-parameter queries can under-report. Final output should not echo Tokens, passwords, or other sensitive credentials.