generated: '2026-09-06'
method: generated
source: >-
  Packaged Agent Skills authored by API Evangelist over the EquipmentWatch OpenAPI harvested
  from https://docs.equipmentwatchapi.com/openapi.yaml on 2026-09-06. EquipmentWatch publishes
  no skills, AGENTS.md or agent documentation of its own — searched 2026-09-06, none found.
provider: Equipmentwatch
providerId: equipmentwatch
description: >-
  Agent Skills for the three marquee EquipmentWatch flows: valuing an asset, verifying a serial
  number, and pricing cost recovery. Each is grounded in real published paths from the OpenAPI.
  IMPORTANT: the source spec declares no operationIds on any of its 24 operations, so every
  skill references the real METHOD and PATH instead. An operationId in these skills would have
  been invented; a path is what the provider actually published.

skills:
  - name: equipmentwatch-value-an-asset
    file: equipmentwatch-value-an-asset.md
    api: equipmentwatch:values
    summary: >-
      Resolve manufacturer and model to a modelRdbId, then retrieve FMV / FLV / OLV for a year,
      condition and region, and optionally trend it.
    operations:
      - GET /taxonomy/manufacturers
      - GET /taxonomy/models
      - GET /values/condition
      - GET /values/region
      - GET /values/value
      - GET /values/trending
  - name: equipmentwatch-verify-serial-number
    file: equipmentwatch-verify-serial-number.md
    api: equipmentwatch:verification
    summary: >-
      Verify year of manufacture from a serial number, surface the serial range behind the
      answer, reconcile against the counterparty's claim, and re-value on the verified facts.
    operations:
      - GET /verification/serialnumberverification
      - GET /taxonomy/models
      - GET /values/value
  - name: equipmentwatch-cost-recovery-rate
    file: equipmentwatch-cost-recovery-rate.md
    api: equipmentwatch:costs
    summary: >-
      Select a build configuration, pull Blue Book ownership and operating cost recovery rates
      plus the FHWA rate, derive an internal charge rate, and compare against retail rental.
    operations:
      - GET /taxonomy/models
      - GET /cost/configurations
      - GET /cost/cost-recovery
      - GET /cost/icr
      - GET /rental/rentalrates

shared_rules:
  - Base URL https://equipmentwatchapi.com/v1; sandbox https://sandbox.equipmentwatchapi.com/v1.
  - Every request carries x-api-key. Keys are sales-issued; there is no self-service signup.
  - Pagination is offset/limit with a hard limit of 50 and no total count in the response.
  - Error envelope is {errorCode, errorMessage, errorDescription} — not RFC 9457. Unmatched
    routes return HTML, not JSON.
  - No rate-limit headers are emitted. Self-throttle.
  - The published API is entirely read-only, so idempotency and reversibility are `na`.
  - No response schemas are published; treat every response field as optional.

not_covered:
  - name: Market Data
    reason: >-
      Documented at https://equipmentwatch.com/api/market-data/ but carries no paths in the
      published OpenAPI. No skill is authored over paths that were never published.
  - name: Integration (saved models and groups)
    reason: >-
      Documented at https://equipmentwatch.com/api/integration/ with POST operations, but absent
      from the published OpenAPI. It is the only write surface EquipmentWatch describes, and
      writing a skill for undocumented write operations — with no known idempotency or reversal
      semantics — would be the exact fabrication this pipeline forbids.