generated: '2026-08-25'
method: generated
source: >-
  Generated from the operationIds MetricStream publishes in its own Business API reference pages
  under https://assets.metricstream.com/pdf/Developer-Portal/ (read 2026-08-25), plus
  conventions/metricstream-conventions.yml, errors/metricstream-problem-types.yml and
  authentication/metricstream-authentication.yml. MetricStream publishes no AGENTS.md and no
  provider-authored skills; these are API Evangelist packaged skills.
grounding: >-
  Every operationId, method and path referenced in these skills was read verbatim from MetricStream's
  published reference pages — none is invented. What the skills deliberately do NOT specify is request
  BODY field names, because MetricStream publishes no machine-readable spec and the section/groupfield
  layout of each entity is configurable per customer instance.
skill_count: 4
skills:
- name: metricstream-report-issue
  file: metricstream-report-issue.md
  api: metricstream-issues
  operations: 5
- name: metricstream-load-risk-register
  file: metricstream-load-risk-register.md
  api: metricstream-grc-foundation
  operations: 10
- name: metricstream-record-loss-event
  file: metricstream-record-loss-event.md
  api: metricstream-loss-event-management
  operations: 7
- name: metricstream-post-kri-metrics
  file: metricstream-post-kri-metrics.md
  api: metricstream-metrics
  operations: 7
shared_cautions:
- No idempotency key on any of the 204 operations — read before every write.
- No delete, cancel, void, undo or restore operation anywhere on the published surface, and no
  retention or correction window is documented.
- No published rate limits and no 429; self-throttle.
- Errors are bare HTTP statuses (200/400/401/403/404/422); no problem+json, no error-code registry.
- The base URL is per-customer-instance and the authentication scheme is not publicly documented.