Shield Agent AI
Support

Scale support & response targets

On Scale we staff a shared Slack (or email) channel and aim for the first-response times below. This page is product policy for hosted Know Your Agent — how we try to show up when something breaks. It is not a service credit schedule, warranty, or indemnity.

Who gets what

Plan Channel Response targets
Starter Email Best effort on business days
Pro Priority email Faster queue than Starter; no severity clock
Scale Priority email + Slack or dedicated channel; guided setup Severity table below

Severity tiers (Scale)

Clock starts when we acknowledge the ticket in the Scale channel or [email protected]. Targets are first human response, not time-to-fix. Business hours mean Monday–Friday, 09:00–17:00 UTC, excluding company holidays.

Severity Examples First response
Sev-1 Hosted console or API down for your workspace; all operators blocked from SSO login; approvals cannot run and agents are stuck on Hold with no workaround 1 business hour
Sev-2 Major feature broken with a workaround (SCIM sync failing, trail ingest delayed, SSO intermittent for some users) 4 business hours
Sev-3 Partial degradation, incorrect role map after an IdP change, non-blocking console bugs 1 business day
Sev-4 How-to questions, SSO/SCIM setup help, feature requests 2 business days

What Scale includes for access

  • Operator SSO (OIDC or SAML) — one IdP per workspace
  • SCIM 2.0 user and group sync
  • IdP group → console role maps (ADMIN / OPERATOR / READ_ONLY)
  • Console roles: OWNER, ADMIN, OPERATOR, READ_ONLY

SSO settings stay with OWNER. Hold approve and kill need an operator capability — OPERATOR can decide; READ_ONLY cannot. Agents keep their own API keys; directory sync never ALLOWs a tool write.

What this page is not

  • Not uptime credits, liquidated damages, or a signed MSA exhibit
  • Not a claim of SOC 2, ISO, or other certification (ask sales for the current packet)
  • Not coverage for self-hosted OSS / local-free runs — those use GitHub issues and docs
  • Not a substitute for /status when you only need live health

Pricing · Status · Docs · Console