Terminal
Terminal/Skill
Troubleshoots why an operator cannot sign in, see, or edit a Konnect resource, checking region, org, team and role scope.
By Kong
Works with
npx skills add kong/ai-marketplace --skill konnect-access-scopeExplain why an identity can or cannot reach a Konnect surface, see a resource, or mutate it, then identify the narrowest next step.
Default to read-first inspection. Do not propose role or team changes until you have separated resource existence, endpoint selection, authentication, and authorization.
Use this skill for operator or automation access to Konnect itself. Do not use it for downstream API consumer auth, Dev Portal application credentials, or Gateway runtime auth flows.
kong-konnect MCP server first for live inspection when it
is available in the session.kongctl-query for read-only checks, identity inspection, and resource
existence checks from the CLI surface.kongctl-declarative for kongctl
YAML and terraform-konnect for HCL-managed Konnect resources.kong-konnect MCP is not
connected, say so early and continue with CLI or user-provided artifacts.Do not treat declarative config in a repo as proof that a user should be able to see or edit the live resource. Live access always wins over intended state.
Load only the reference file that matches the active branch:
references/region-org-and-team-checks.md
references/visibility-vs-permission.md
references/idp-and-role-model.md
Pin down:
kongctl, Konnect UI, Konnect API, or
Portal operator UIKeep the failing operation concrete. Access failures look similar across surfaces but have different root causes.
If the complaint is really about developer-facing application registration,
issued credentials, or runtime API access, hand off to konnect-app-auth
instead of stretching this skill.
Confirm the caller is pointed at the right Konnect region and organization before reasoning about roles.
For CLI-based checks, prefer:
kongctl get organization -o json
kongctl get me -o json
If those succeed, authentication is present. If they fail, stop there and fix auth or endpoint selection first.
Load references/region-org-and-team-checks.md when the main ambiguity is
where the caller is pointed rather than what permissions they have.
Check whether the target resource exists in the organization independently of the affected user.
Load references/visibility-vs-permission.md when the core question is whether
the resource is absent, hidden, or only non-mutable.
Interpret access using these defaults:
Do not assume a missing mutation permission implies missing read permission, or the reverse.
When the user mentions Okta, SSO, SCIM, or IdP-managed groups, verify whether the requested change is blocked because identity membership is mastered outside Konnect.
Treat externally managed identity data as read-only from the Konnect side unless you can confirm otherwise.
Load references/idp-and-role-model.md when role layering, IdP mastery, or
team/identity mapping is the main decision branch.
Classify the outcome as one of:
Then give the minimum next action needed to unblock the user.
If the root cause is outside Konnect operator access, say which neighboring skill should take over instead of keeping the diagnosis open-ended.
Before answering, verify that you can state:
kongctl-query for exact read-only command selection and output shaping.kongctl-declarative or terraform-konnect when the user wants to
create or update managed Konnect resources after the access issue is
understood. Match the repository's current toolchain.konnect-app-auth when the real issue is developer registration,
application credentials, or Dev Portal auth behavior rather than operator
access to Konnect.konnect-gateway-triage, konnect-observability-triage,
konnect-api-publish, or konnect-event-gateway when the real problem is
not access but operator ambiguity inside that surface.