uipath-admin

UiPath Admin

Administrative operations on UiPath via uip admin — Identity Server, Authorization, OMS, IP Restriction, Audit. Per-area workflows, command references, and procedures are in the linked files below — this file is the entry contract.

When to Use This Skill

Identity

• Manage identity users — list, create, invite, update, delete
• Manage groups — CRUD + add/remove members
• Manage robot accounts — create, update, delete unattended robot identities
• Manage external apps — OAuth2 clients, secrets, federated credentials
• Manage personal access tokens (PATs) — create, list, revoke, regenerate
• Configure SMTP — get, update, test, delete email settings
• Browse OAuth2 scopes — list available scopes for external apps and PATs
• Onboard human user — invite, assign to groups
• Onboard robot account — create account, assign to groups

Authz

• Manage custom roles — CRUD on Authorization service role definitions (scope shapes: Organization, TenantGlobal, Tenant, Project)
• Manage role assignments — assign roles to users/groups/robot accounts at Organization, Tenant, TenantGlobal, Project, Folder, or App scope
• List permission definitions — read-only catalog of permissions across services
• Check effective access — compute what a principal can actually do at a given scope (Policy Decision Point)
• Grant permission(s) to a principal — ad-hoc "grant me X" / "give Y, Z" requests resolved via the scope/service intersection flow

OMS

• Inspect / update the current organization — uip admin organizations (read + update only; no CLI create/delete)
• Manage tenant lifecycle — create, enable, disable, delete tenants in the caller's org
• Provision org-level or tenant-level services — services list, list-available, add, enable, disable, remove
• Poll async OMS operations — tenants mutations return operationId; poll via organizations operation get <id> (the canonical poll endpoint)
• List available regions — discover provisioning regions before tenants create

IP Restriction

• Manage IP allowlisting — add / update / delete CIDR entries that gate inbound access
• Toggle IP-restriction enforcement — turn the org-wide allowlist switch on or off (with lockout safety)
• Manage bypass rules — URL-pattern exceptions to IP allowlisting
• Look up the caller's public IP — answer "what's my IP?" / "what public IP does the platform see for me?" directly with ip-restriction my-ip; also the safety pre-flight before enabling enforcement

Audit

Activate on both explicit audit requests and natural-language investigation intent — users rarely say "audit events" by name.

Surface boundary — this is uip admin audit, not uip or audit-logs. Every org/tenant audit ask below — including bare "audit logs", "audit history", "export the audit trail", "login history", "who did what" — is served by uip admin audit <scope>. Do not fall back to uip or audit-logs (Orchestrator-operational audit, a different schema; that's the uipath-platform skill), even when the user's wording is generic. When the user asks what audit events / sources are available or visible, answer by running uip admin audit <scope> sources — discover live, never describe from memory or the docs.

• Explicit — uip admin audit commands; list sources / targets / types; query, filter, paginate, or export events; CSV or per-day-JSON dump of audit history for a window.
• Query audit events — list event sources, filter events by source / target / type / user / status / time window at org or tenant scope
• Export audit events — chunked download from the long-term store (one call per UTC day, atomic abort on any chunk failure) as a folder of day-wise JSON files (default) or a single merged CSV via --file-format csv
• Membership / license phrasings — "who joined / left the organization", "who was made an admin", "license changes", "cross-tenant audit"
• Sign-in / authentication phrasings — "failed/successful logins", "login history for user X", "who's been signing in"
• Tenant-activity phrasings — "what happened on tenant X", "asset/queue/folder edits", "queue items processed", "job failures", "Action Center task changes", "Apps / AgentHub / Document Understanding / Integration Service / Test Manager activity"
• Cross-scope phrasings — "everything everywhere" (run the flow once per scope and present combined)
• Investigation intent (full-sentence form) — "Who deleted the X folder last Tuesday?", "Show me failed logins for user Y this month.", "What changed on tenant Z between Jan 1 and Feb 1?", "Give me the audit log for the last 30 days.", "Was the API key rotated by someone in our org?", "Export everything for compliance for Q4."

Scope routing (which phrasing → org vs tenant, and why) lives in audit-workflow-guide.md → Audit scope disambiguation. Critical Rule 23 governs the stop-and-ask requirement when scope is ambiguous.

Troubleshoot

Activate on access/auth/identity troubleshooting — users report symptoms, not audit verbs.

• Diagnose access denied — "user can't access X", "403 on API", "new hire has no permissions" → resolve principal, check-access, inspect role assignments. Playbook: identity-troubleshoot-guide.md → Playbook 1
• Investigate login failures — "failed login attempts", "account compromised?", "suspicious sign-ins" → org-scoped audit login-history investigation. Playbook: identity-troubleshoot-guide.md → Playbook 2
• Diagnose role misconfiguration — "custom role doesn't work", "user has role but can't do X" → inspect role actions, verify scope alignment. Playbook: identity-troubleshoot-guide.md → Playbook 3
• Diagnose IP restriction lockout — "can't access platform from new office", "all users blocked" → my-ip + ip-ranges list + enforcement get. Playbook: identity-troubleshoot-guide.md → Playbook 4
• Diagnose PAT / external app failures — "API returns 401", "PAT stopped working", "external app can't authenticate" → check expiry, scopes, audit for revocation. Playbook: identity-troubleshoot-guide.md → Playbook 5
• Diagnose SMTP email delivery failures — "invitations not sending", "SMTP broken" → smtp get + smtp test
• Investigate stuck tenant operations — "tenant create not completing", "operation stuck" → poll operation status
• Identify service provisioning no-ops — "service still enabled after remove" → platform-pinned services
• Triage robot account authentication issues — "robot not authenticating" → identity vs credential model confusion

Structured diagnose capability index with failure-mode lookup and diagnostic priority ladder: diagnose/CAPABILITY.md. Quick investigation playbooks: identity-troubleshoot-guide.md.

Critical Rules

Each rule is the agent contract. Per-area detail is in the linked reference files.

Universal

1. Route Orchestrator-specific role/permission requests to uip or roles (uipath-platform skill). uip admin authorization does NOT own Orchestrator's role catalog. Conversely, all organization/tenant audit is owned HERE and served by uip admin audit <scope> (sources / events / export) — NEVER use uip or audit-logs for audit events, audit history, audit export, login history, compliance dumps, or "who did what/where" requests. uip or audit-logs is a different surface (Orchestrator-operational audit — Component,User,Action,Operation,Time columns, --export returns a CSV; the uipath-platform skill). If you reach for uip or audit-logs on any org/tenant audit ask, stop and switch to uip admin audit.
2. Verify login first. uip login status --output json. If not logged in, stop and ask the user to run uip login — it opens an interactive browser flow, so never launch it yourself in an automated or non-interactive session (sessions authenticated via env vars are already logged in). Org id is resolved from the active session.
3. Use --output json on every command. Parse programmatically; present conversationally.
4. Stop on error. Show the error verbatim. Never retry auth failures — ask the user to uip login.
5. Resolve every named principal before high-risk ops. Any command that touches a named user / group / robot account / external app — roles assignments create/delete, users delete, groups delete, groups members add/revoke, robot-accounts delete, external-apps delete, external-apps generate-secret — MUST first search the directory and echo Principal: <displayName> (<userName>) — <id> back before the mutation runs. Zero matches → stop and ask; never fall back to the current login user. Multiple matches → numbered list, wait for a digit. Procedure: role-assignment-management.md → Resolving Principal IDs.

Identity

6. Discover before creating. list before create to avoid duplicates (robot accounts, groups, external apps — users invite excepted).
7. Secrets shown only once on external-app create and generate-secret — warn the user to save immediately.
8. External apps require scopes at creation — --app-scope or --user-scope is required (e.g., --app-scope "OR.Folders").
9. Group membership uses user IDs. Resolve via users list per Rule 5, then groups members add/revoke.
10. Confirm before delete on users / groups / robot accounts / external apps — after resolving the named target per Rule 5.

Authz

11. Built-in roles are read-only. Only Custom roles can be created / updated / deleted. CLI also rejects authoring against service-managed and platform-level services. Service lists: role-management.md → Services That Manage Their Own Roles.
12. roles create / roles update are PUT-style upserts. Body is assembled from inline flags + --file ./actions.json. Always roles get first before updating — omitted flags overwrite that field.
13. --service infers scope (e.g., --service studio → Tenant; --service apps → Organization). Combine with --scope only to override. Never guess a serviceName — the valid values and the command that re-derives them: permission-catalog.md → --service serviceNames.
14. Listing works for every service; authoring is what's blocked. roles list --service <svc> and roles assignments list --service <svc> accept every service. For effective access on a principal use check-access (PDP).
15. Scope vocab differs across verbs. roles create --scope: Organization|TenantGlobal|Tenant|Project. roles assignments create --scope: those + Folder|App. roles assignments list --scope: excludes TenantGlobal. check-access --scope: only Tenant|Folder.
16. roles assignments create/delete MUST resolve the principal first per Rule 5 — --identity-id is a raw UUID the CLI does not name-check.
17. roles assignments create MUST match the role's ownerServiceName to the scope-path service segment. CentralizedAccess → no service segment (/ or /tenant/<tid>); anything else → path must include lowercase(ownerServiceName). Display-name mapping (e.g., Reinfer → "IXP") + full procedure: role-assignment-management.md → Validate Role's Owning Service.

OMS

18. Async lifecycle: auto-poll, then hand off. tenants create/update/delete/enable/disable return operationId. Auto-poll organizations operation get <OP_ID> 3× at 5 s; on terminal status stop and report; still in-progress after 3 polls → numbered menu, never indefinite loop. organizations create and organizations delete are not exposed by the CLI — Portal / support flow only. Procedure: organization-management.md → Polling procedure.
19. tenants delete is soft-only. No hard-delete flag; restoration is via support.
20. Tenant commands default to the login tenant. Always pass an explicit <TENANT_ID> for destructive ops (tenants delete, tenants disable, tenants services remove).
21. Resolve region before tenant create. --region is required on tenants create — run organizations regions list first. Tenant service catalog is region-aware.
22. services disable / remove may no-op despite Success on certain services. Always re-list after mutating. Gap list: tenants-commands.md → Concepts.

Audit

23. Disambiguate org vs tenant scope before querying. If the prompt is vague AND no prior turn fixed the scope, stop and ask once (a single clarifying question — use AskUserQuestion when available) — never silently default to tenant or any single scope. If you genuinely cannot get an answer (non-interactive run) and must proceed, query both scopes and combine the results rather than guessing one — silently picking a single scope is the failure mode. Scope is a positional subgroup, NOT a flag: write uip admin audit org sources / uip admin audit tenant events — there is no --scope option (audit sources --scope organization is invalid). Routing table (user-phrasing → scope + why it lives where) and investigation playbooks: audit-workflow-guide.md → Audit scope disambiguation.
24. audit <scope> events returns an object, not a bare array. Shape is {auditEvents, next, previous}. Do not index Data[0]; read Data.auditEvents[]. Cursor semantics are chronological: next = newer events, previous = older events. The default newest-backward walk follows previous.
25. --limit paginates internally — never loop on --from-date / --to-date to "paginate". The server clamps maxCount to [10, 200] per request; when the user wants more than 200, the CLI fetches ceil(N/200) pages under the hood. Pass --limit 500 (or larger, up to the 10000 maximum) — do NOT re-implement pagination in the agent. --limit must be in [1, 10000]; a bigger value is rejected up front with Result: "ValidationError", so never pass an arbitrarily huge number to mean "everything" — for that, omit --limit or stay within the range.
26. Discover via audit <scope> sources first — never invent source / target / type GUIDs. The catalog response gives the GUIDs you pass to events --source / --target / --type. This is also the answer to "what audit events / sources can I see?" — run audit <scope> sources and report the live catalog; do not answer conceptually from memory or the reference docs.
27. Bound the time window, ISO 8601 in UTC. Don't call audit <scope> events without --from-date and --to-date on a noisy tenant. Accepted formats: date-only (2026-04-01) or with time (2026-04-01T14:30:00Z). --to-date is inclusive of the exact instant — to capture a full final day, pass the start of the next day or T23:59:59.999Z. This next-day trick is for events only — export bounds are whole days (Rule 30). Resolve relative phrases ("yesterday", "last 7 days") against the actual current UTC date (e.g. date -u) — never guess dates — and echo the resolved window in your reply. 27b. "No matching event" is a complete answer — never name an actor the evidence does not support. An audit trail is evidence, so attributing an action to the wrong person is worse than returning nothing. If the targeted query comes back empty, say so plainly, state what you searched (scope, source/target/type, window), and offer a next step (widen the window, try the other scope, check the resource ever existed). Never promote an actor from an adjacent event — a different resource, a different event type, or a broad --search hit — into the answer, and do not keep loosening filters until something returns and then present that as the culprit. Only name an actor when the cited event matches both the resource and the verb asked about, quoting createdOn plus the identifying eventDetails field so the user can verify it. Detail: audit-workflow-guide.md → Step 5.
28. --tenant-id is silently ignored on org-scoped audit commands. If you find yourself reaching for it on audit org events, switch to audit tenant instead.
29. On 401 from audit, do NOT retry. The token is missing the Audit.Read scope; tell the user to uip logout && uip login. 29b. On a 5xx from audit, DO retry — the envelope tells you so. ErrorCode: server_error / Retry: RetryLater (e.g. HTTP 503: no healthy upstream, HTTP 504) is a transient, not a bad request. Re-run the same query up to two more times with a few seconds of backoff before reporting failure — changing --limit or the window between attempts is not a fix and only obscures the cause. Retry immediately-and-once is usually not enough; wait between attempts. Never write or present an error envelope as if it were the result — if the user asked you to save the JSON, saving {"Result":"Failure",...} to that file is a failed retrieval, not a result. Report the failure explicitly instead.
30. audit <scope> export writes into a base directory (--output-path): a uniquely-named folder of day-wise JSON files (default) or a single merged CSV. --from-date, --to-date, and --output-path are all required. Export dates are whole UTC days, inclusive on both ends (the server truncates times to the calendar day): --from-date 2026-01-01 --to-date 2026-01-31 exports all of January; a single day is --from-date X --to-date X with the same date. Do not apply the events next-day trick here — a next-day --to-date exports an extra full day. --output-path is a base directory (created if missing) — pass a directory only, never a filename or extension; the CLI creates a uniquely-named audit_<from>_<to>_<generated-at> output inside it (folder for json, .csv for csv), so repeated exports of the same window never collide. Do not hand-craft the per-export name. --file-format <json|csv> selects the shape: json (default) = a folder of <YYYY-MM-DD>.json files; csv = one merged CSV — pick csv for a flat spreadsheet/Excel-friendly dump, json for per-day files. If the user named a destination folder, pass it verbatim as --output-path — no confirmation needed; confirm first only when you had to pick a default (e.g. ./audit-exports) yourself. Then report the generated Path (and GeneratedAt) from the result.

IP Restriction

31. enforcement enable is lockout-sensitive — prompt + impact statement required. Run ip-restriction my-ip and verify the caller's IP is covered by an entry in ip-ranges list. Then prompt the user with the impact before flipping: "After enabling IP restriction, any caller (Portal, CLI, robot, external app) whose source IP is not in ip-ranges list will be blocked from this org. Misconfiguration locks you out and requires platform-side recovery. Proceed?" --confirm is required; ip-ranges delete while enforcement is on also requires --confirm. Procedure: enforcement-management.md.
32. Recovery from IP lockout requires platform-side action. No CLI bypass — either access from an in-allowlist IP and enforcement disable, or use the Portal recovery flow.
33. "APMS" is internal — never expose to the user. "APMS" (Access Policy Management Service) is the platform's internal name for IP Restriction. Use "IP Restriction" in every user-facing surface.

What NOT to Do

1. Never delete built-in groups. type: "BuiltIn" groups cannot be deleted. Only custom groups.
2. Never pass IDs as flags. Resource IDs and names are positional arguments: groups members add <GROUP_ID> --user-ids ..., NOT --group-id <GROUP_ID>. Same for all get, update, delete, create subcommands.
3. Do NOT assume audit events returns a bare array. It's {auditEvents, next, previous}.
4. Do NOT loop on --from-date/--to-date to "paginate". Bump --limit and the CLI handles cursor pagination internally.
5. Do NOT silently default audit scope to tenant or org when the prompt is ambiguous. Ask once, then proceed.
6. Do NOT invent audit source/target/type GUIDs. Always discover via sources first.
7. Do NOT call audit events with no time bound on a noisy tenant — default to a bounded window.
8. Do NOT pass --tenant-id to org-scoped audit commands — it's silently ignored.
9. Do NOT retry on 401 auth errors. The token is missing the required scope (Audit.Read for audit). Tell the user to uip logout && uip login. 9a. Do NOT save or report an audit error envelope as data. A {"Result":"Failure"|"ValidationError", ...} payload is a failed call — retry a 5xx (Rule 29b), fix a ValidationError, and never leave it in a file the user asked you to fill with results. 9b. Do NOT name a culprit an audit query did not return. If the targeted search is empty, report "no matching event found" — never borrow an actor from an adjacent or broader-search event (Rule 27b).
10. Do NOT call roles update with only the flag you want to change. Re-fetch first; the upsert body overwrites omitted fields (Rule 12).
11. Do NOT present authz results without provenance — role name, scopeType, ownerServiceName, tenant-binding (names not UUIDs). Detail: authorization-commands.md → Provenance contract.
12. Do NOT conflate provisioned services with the available catalog. services list returns provisioned with status; services list-available is the catalog. Present them as separate sections.
13. Do NOT run an OMS mutation without naming the target. Echo org name / tenant name + UUID / service type + region before running.
14. Do NOT pass the next day as --to-date on audit export to "include the final day" — export bounds are already whole-day inclusive (Rule 30); the next-day trick is for events only and would export an extra day.

Quick Start

One row per common goal. Per-area workflows are in the reference files.

Key Concepts

Organization hierarchy

Organization (org)
  └── Partition (= org in most cases)
        ├── Users           ← human identities
        ├── Groups          ← role containers (BuiltIn + Custom)
        ├── Robot Accounts  ← unattended automation identities
        └── External Apps   ← OAuth2 clients (Client ID + Secret)

Robot accounts vs external apps

Robot credentials are provisioned automatically by Orchestrator on machine connect — not by creating external apps.

Output Etiquette

What to surface after each verb. Per-area detail in the reference files; this is the contract.

For per-area full checklists, follow the table's inline links: Identity → identity-commands.md; Authz → authorization-commands.md; Audit → audit-workflow-guide.md.

Task Navigation