Account integrations (Studio connected accounts)

Account integrations are the account-scoped (Studio) connected accounts — Gmail, Slack, Salesforce, Google Sheets, etc. Each one has a numeric id that is the connectedAccountId you wire into workflow/flow triggers, tool nodes, and the resource pickers below. Setting up an automation almost always starts here: find the integration, confirm it’s connected, and read its id.

All frontline integrations commands are USER-scoped — a GENERAL/account key returns 401. Auth uses the same per-user API key as the rest of the CLI (frontline auth login <key>).

Managing integrations

integrations list filters: --toolkit <TOOLKIT>, --filter-text <text>, --connected, --disconnected. Toolkit values are UPPERCASE enum names (GMAIL, SLACK, SALESFORCE, GOOGLESHEETS, HUBSPOT, …) — integrations toolkits is the authoritative list; an unknown value returns 400. Add --table to any read command for a compact view.

# Find the Gmail integration and grab its id (the connectedAccountId for triggers).
frontline integrations list --toolkit GMAIL --connected --table
#   → id: 42   toolkit: GMAIL   alias: Support inbox   isConnected: true

# See what a toolkit can do, then enable a specific set of tools.
frontline integrations toolkits --table
frontline integrations tools list 42 --table
frontline integrations tools set 42 --slugs GMAIL_SEND_EMAIL,GMAIL_FETCH_EMAILS

Output shapes (default JSON)

• integrations list → { "results": [ … ] } (no pagination). Key fields per integration: id (the connectedAccountId), toolkit, alias, visibility, isConnected, isExpired, connectedEmail, allowedToolSlugs (currently enabled), enabledToolsCount, availableToolsCount, createdByUser.
• integrations toolkits → { "results": [ { toolkit, name, description, authScheme, toolCount, tools[] } ] }. tools[] are the valid slugs for tools set.
• integrations tools list <id> → { "enabledTools": [...], "availableTools": [...] }, each entry { slug, name, description, enabled, hasRequiredScopes }. hasRequiredScopes: false means the connected account’s OAuth grant is missing the scopes that tool needs — it will fail at runtime until the user reconnects the integration in the Studio UI with broader permissions. Don’t enable such tools without flagging this.

Using the id in workflows and flows

This is what the id is for (see the workflow-builder skill for full node payloads):

• Event trigger node: a workflow TRIGGER node with triggerType: "COMPOSIO_TRIGGER" — its data requires connectedAccountId, triggerToolkit (e.g. GMAIL), and triggerSlug, plus the trigger’s triggerConfig (whose IDs you resolve with the resource pickers below). Discover valid slugs + their config fields with max tasks triggers.

frontline workflows nodes create --data '{"nodeId":"node_11111111-1111-1111-1111-111111111111","type":"TRIGGER","position":{"positionX":0,"positionY":0},"data":{"type":"TRIGGER","triggerType":"COMPOSIO_TRIGGER","connectedAccountId":42,"triggerToolkit":"GMAIL","triggerSlug":"GMAIL_NEW_GMAIL_MESSAGE","triggerConfig":{"labelIds":"INBOX"}}}'

When the trigger fires, the event payload is available in downstream nodes as the built-in {trigger_payload} variable (plain string, no dot-access — extract fields with DATA_TRANSFORMER or AI_CAPTURE, same as {webhook_payload}).

• AI tool nodes (TOOLS_AI): node data.connectedAccountConfigs is an array of { connectedAccountId, useAllTools, selectedToolSlugs } — so a node can use all of the integration’s enabled tools ("useAllTools": true) or a narrower subset ("useAllTools": false, "selectedToolSlugs": ["GMAIL_SEND_EMAIL"]).
• These references are validated server-side: a wrong connectedAccountId returns 400 bad_request with details.issues[].code === "not_found" and the node is dropped.

Semantics, permissions, and errors

• tools set is a full replace, not a merge — pass every slug you want enabled. Slugs must belong to the integration’s toolkit (invalid slugs → 400 “Invalid tool slugs: …”). The change is account-wide: every workflow/agent node using this integration with useAllTools: true immediately gains/loses those tools.
• Visibility: SHARED = usable by the whole account; PRIVATE = only the creator. integrations list only returns integrations visible to the calling user, so a missing integration may simply be someone else’s PRIVATE one.
• Only the creator, an admin, or an owner may update an integration or change its tools — otherwise 403. Renaming to an alias that already exists in the account returns 409.
• Connecting a new integration (OAuth) is not available from the CLI — it needs an interactive browser redirect. Connect it in the Studio UI, then manage it here. Disconnecting/deleting is also UI-only.
• An integration that is isConnected: true but isExpired: true needs to be reconnected in the UI before its tools/triggers will run.
• A GENERAL/account API key returns 401 on every command in this skill; a wrong integrationId returns 404.

Integration resources (trigger ID resolver)

Event triggers don’t take human names — they take IDs. A “new rows in a Google Sheet” trigger needs a spreadsheet_id, not “my budget sheet”. These commands list the resources in a user’s connected integration so you can pick the right ID and drop it into a trigger config.

This is the missing middle step of the event-trigger flow:

1. Discover the trigger — max tasks triggers (or the workflow trigger catalog): gives you the slug and its config field names (e.g. spreadsheet_id, sheet_name).
2. Resolve the IDs — these commands: turn names into IDs for those config fields.
3. Create — pass the resolved IDs in the trigger config (max tasks create --trigger-slug … --trigger-config '{…}', or the workflow trigger node).

If a picker returns nothing or 401s unexpectedly, check integrations list first — the toolkit may not be connected, may be expired, or may be someone else’s PRIVATE integration.

Which command resolves which config field

The searchable pickers (sheets, docs, drive folders/drives) accept --query <text> and --limit <n> (1–100, default 50). Add --table for a readable list.

Typical flow

# 1. See what the trigger needs.
max tasks triggers --pretty
#   → GOOGLESHEETS_NEW_ROWS_TRIGGER needs { spreadsheet_id, sheet_name, start_row }

# 2. Resolve the spreadsheet, then its tabs.
frontline integrations resources googlesheets spreadsheets --query "Q3 budget" --table
#   → id: 1AbC...   name: Q3 Budget
frontline integrations resources googlesheets sheets 1AbC... --table
#   → title: Sheet1

# 3. Create the task with the resolved IDs.
max tasks create \
  --name "New budget rows" \
  --instructions "Summarize each new budget row and flag anything over $10k." \
  --trigger-slug GOOGLESHEETS_NEW_ROWS_TRIGGER \
  --trigger-config '{"spreadsheet_id":"1AbC...","sheet_name":"Sheet1","start_row":2}'

Multiple accounts of the same toolkit

If the user has more than one connected account for a toolkit (e.g. two Gmail accounts), pass --connected-account-id <id> to target a specific one — the <id> is the integration id from integrations list. With a single account per toolkit you can omit it.

Examples

frontline integrations resources gmail labels --table
frontline integrations resources googledocs documents --query proposal
frontline integrations resources googledrive folders --query Invoices --table
frontline integrations resources asana workspaces
frontline integrations resources asana projects 1200000000000001
frontline integrations resources salesforce sobjects --table
frontline integrations resources salesforce fields Opportunity
frontline integrations resources outlook calendars

See also

• Building workflows/flows (exact TRIGGER and TOOLS_AI node payloads where connectedAccountId goes): the workflow-builder skill.
• Max automated tasks (where the resolved resource IDs get used): the max-tasks skill.
• Full reference: https://docs.getfrontline.ai/cli.