picked 0.0.101-next.2 → 0.0.101

package/docs/reference/dashboard/lead-search.md

@@ -0,0 +1,235 @@
+---
+type: reference
+layout: docs
+---
+
+# Lead search
+
+A lead is one set of answers that a visitor gave to your flow. The **Leads** page lists
+them, and it has one search box. The box accepts a plain word, a filter, or both. The
+dashboard holds the query in the address, so you can share the link.
+
+## The parts of a lead that a query reads
+
+| Part | How to write it | Example |
+| --- | --- | --- |
+| Every captured answer | a plain word | `melbourne` |
+| One label | `namespace.key` | `acme.status:sent` |
+| One answer | the key of the field | `suburb:Sydney` |
+| The capture date | `createdAt` | `createdAt>2026-08-01` |
+| The visitor | `userId` | `userId:sess_7x2k9q3m1n4a` |
+| The visit that captured the lead | `session.field` | `session.utm_source:google` |
+
+A path of two parts is a label, unless it starts with `session.`: that path reads the visit.
+A label namespace and a label key cannot contain a full stop. A path of one part is the key
+of a field, unless the name is `createdAt` or `userId`.
+
+A query does not read an expired label. Refer to [Labels](/reference/dashboard/labels#expiry).
+
+## Operators
+
+| Operator | Meaning | Example |
+| --- | --- | --- |
+| `:` | equal to | `details.phone:"0400000000"` |
+| `~` | contains | `details.email~example.com` |
+| `:*` | the lead has a value | `acme.status:*` |
+| `>` `>=` `<` `<=` | compare a number or a date | `scoring.value>=50` |
+| `-` | makes the clause negative | `-acme.status:*` |
+
+Each part accepts a different set:
+
+| Part | `:` | `~` | `:*` | `>` `>=` `<` `<=` |
+| --- | --- | --- | --- | --- |
+| Label | yes | yes | yes | yes |
+| Answer | yes | yes | yes | no |
+| `createdAt` | yes | no | no | yes |
+| `userId` | yes | no | no | no |
+| `session.` | yes | yes | yes | no |
+
+## How clauses combine
+
+A space between two clauses means AND. The word `OR` between them means OR.
+
+```
+acme.status:sent scoring.contacted:true
+acme.status:sent OR acme.status:queued
+```
+
+One query holds a maximum of 10 clauses. One query uses AND or OR, and not both: there is
+no way to show which one comes first. A query that mixes them gives an error.
+
+## The visit that captured the lead
+
+A visit is the set of events that one visitor made, from the first page to the last page. The visit
+that captured a lead is the visit that holds its capture. A clause on the visit matches the
+lead when one event of the visit has the value. The event can come before the capture or
+after it.
+
+| Field | What it holds |
+| --- | --- |
+| `gaid` | The ID that Google Analytics gives the visitor |
+| `utm_source` | The `utm_source` parameter of the page address |
+| `utm_medium` | The `utm_medium` parameter of the page address |
+| `utm_campaign` | The `utm_campaign` parameter of the page address |
+| `utm_term` | The `utm_term` parameter of the page address |
+| `utm_content` | The `utm_content` parameter of the page address |
+| `page_url` | The full address of the page |
+| `page_origin` | The host of the page, for example `shop.example.com` |
+| `page_pathname` | The path of the page, for example `/quote` |
+| `page_query` | The query string of the page, with no `?` |
+| `landing_page_url` | The full address of the first page of the visit |
+| `landing_page_origin` | The host of the first page, for example `shop.example.com` |
+| `landing_page_pathname` | The path of the first page, for example `/` |
+| `landing_page_query` | The query string of the first page, with no `?` |
+
+A clause on the visit searches every lead, with no date limit.
+
+A clause with `-` matches a lead when no event of the visit has the value. A lead that has
+no visit does not match a clause on the visit, with `-` or without it. A lead has no visit
+when one of these is true:
+
+- The lead has no capture event. The capture came before Picked recorded capture events, or
+  the deployment sends no events.
+- The visit was on a deployment that is not for production, for example a preview or a test
+  deployment.
+- The visit started more than 7 days before its events arrived.
+- The time on the device of the visitor was more than one day later than the real time.
+
+A query with a clause on the visit combines its clauses with AND. It can use OR only when
+all of its clauses are on the visit.
+
+A query with a clause on the visit shows the leads in the order of their capture events,
+newest first. A capture event has its own time, and that time can differ from the `createdAt`
+of the lead. The dashboard shows no column sort while the search has a clause on the visit.
+
+A query with a clause on the visit shows the number of leads on the page, and no total. For
+a page of 20 leads with another page after it, the dashboard shows `Showing 20 leads` above
+the list.
+
+A clause with `:` finds a value that is exactly the same, with the same upper case and lower
+case.
+
+In a clause on the visit, `~` finds the value as text, and ignores the difference between
+upper case and lower case. In a clause on the visit, `%` and `_` in the value match only `%`
+and `_`.
+
+## Values with a space
+
+Put `"` at each end of a value that has a space. To search for `"` inside a value, write
+`\"`.
+
+```
+details.company:"Acme Corporation"
+```
+
+## Plain words
+
+A plain word searches every answer that the visitor gave, and it includes the text inside
+a JSON answer.
+
+A plain word matches a whole word. `amy@example.com` finds the lead with that address, and
+`example.com` finds nothing. To match a part of a value, name the field and use `~`:
+
+```
+email-address~example.com
+```
+
+Two plain words both must match.
+
+## Examples
+
+**Plain words**
+
+| Query | What it finds |
+| --- | --- |
+| `melbourne` | Leads with `melbourne` in an answer |
+| `amy@example.com` | The lead with that address |
+| `Amy Rockwell` | Leads with both words in their answers |
+| `dental` | Leads with `dental` inside a JSON answer |
+
+**Labels**
+
+| Query | What it finds |
+| --- | --- |
+| `details.phone:"0400000000"` | Leads with that exact telephone number |
+| `details.email~example.com` | Leads with an email label that holds `example.com` |
+| `acme.status:sent` | Leads that the `acme` workflow sent |
+| `acme.status:*` | Leads with an `acme` status |
+| `-acme.status:*` | Leads with no `acme` status, which no workflow took |
+| `scoring.value>=50` | Leads with a score of 50 or more |
+| `scoring.contacted:true` | Leads with that boolean label set to true |
+| `crm.payload:*` | Leads with a JSON label under `crm` |
+
+**Answers and the capture date**
+
+| Query | What it finds |
+| --- | --- |
+| `suburb:Sydney` | Leads whose `suburb` answer is `Sydney` |
+| `email-address~rocketrides` | Leads whose `email-address` answer holds `rocketrides` |
+| `quote-preferences:*` | Leads that answered `quote-preferences` |
+| `createdAt>2026-08-01` | Leads captured after that date |
+| `userId:sess_7x2k9q3m1n4a` | Leads captured for that visitor |
+
+**The visit that captured the lead**
+
+| Query | What it finds |
+| --- | --- |
+| `session.utm_source:google` | Leads whose visit had the `utm_source` `google` |
+| `session.utm_campaign~spring` | Leads whose visit had a `utm_campaign` that holds `spring` |
+| `session.page_pathname:/quote` | Leads whose visit had the page `/quote` |
+| `session.page_origin:shop.example.com` | Leads whose visit had a page on `shop.example.com` |
+| `session.landing_page_pathname:/` | Leads whose visit started on the page `/` |
+| `session.landing_page_query~utm_source=google` | Leads whose visit started on a page with `utm_source=google` in its query string |
+| `session.gaid:*` | Leads whose visit had a Google Analytics ID |
+| `-session.utm_source:*` | Leads whose visit had no `utm_source` |
+| `session.utm_source:google OR session.utm_source:bing` | Leads whose visit had either `utm_source` |
+
+**More than one clause**
+
+| Query | What it finds |
+| --- | --- |
+| `details.first_name:Amy scoring.contacted:true` | Both clauses are true |
+| `acme.status:sent OR acme.status:queued` | Either clause is true |
+| `Sydney details.phone:"0400000000"` | A plain word and a label clause together |
+
+## The `labels.` and `responses.` prefixes
+
+A path takes an optional `labels.` or `responses.` prefix. Two cases need it:
+
+- A label namespace with the name `labels`, `responses`, or `session`: write
+  `labels.responses.status:sent`.
+- The key of a field that holds a full stop: write `responses.quote.preferences:*`. Without
+  the prefix the query reads that path as a label.
+
+A field with the name `createdAt` or `userId` also needs the prefix. A path of one part with
+that name reads as the capture date or the visitor.
+
+## Errors
+
+The search box shows an error, and the list does not change, for each of these:
+
+| Cause | Example |
+| --- | --- |
+| A `"` with no second `"` | `details.company:"Acme` |
+| A namespace or a key outside `a-z`, `0-9`, `_` and `-` | `Details.Phone:1` |
+| An operator that the part does not accept | `createdAt~2026` |
+| AND together with OR | `a.b:1 c.d:2 OR e.f:3` |
+| More than 10 clauses | 11 clauses in one query |
+| A visit field that does not exist | `session.referrer:x` |
+| OR between a clause on the visit and a clause that is not | `session.utm_source:google OR labels.crm.status:sent` |
+
+## API and MCP
+
+You can also do this task with the [Management API](/reference/management-api/schema), the [MCP server](/reference/tools/mcp), or code that you send to [`picked_run`](/reference/tools/mcp-run).
+
+| Task | GraphQL | MCP tool | Run |
+| --- | --- | --- | --- |
+| Search the leads | [`project`](/reference/management-api/schema/project#project) | None | None |
+
+The `leads` field of the `project` query takes the same search text in its `query`
+argument.
+
+## Related
+
+- [Fields](/reference/dashboard/fields) — the keys that name each answer
+- [Glossary](/reference/glossary) — lead, label, and the other terms

package/docs/reference/dashboard/projects.mdx

@@ -0,0 +1,53 @@
+---
+type: reference
+layout: docs
+---
+
+import { Screenshot } from '../../../../components/Screenshot'
+
+# Projects
+
+A project is one site or one application that you want to measure with Picked. Each
+project has its own experiences, content, and options.
+
+<Screenshot
+  route="/{orgSlug}"
+  alt="The organization home page on the Picked dashboard, with the projects in an organization"
+/>
+
+## Workspaces
+
+Each project holds one or more [Workspaces](/reference/dashboard/workspaces). In a workspace you set the
+options for an experience and you see a preview of it before you make it live.
+
+## Project home
+
+When you open a project, you land on the project home page. It shows your SDK
+snippet and a list of the setup tasks that are still open.
+
+Copy the snippet and add it to your site. The snippet connects your site to the
+project and lets Picked serve experiences to it. See
+[SDK Installation](/reference/sdk) for the setup detail.
+
+## Project navigation
+
+The sidebar shows the sections of your project: **Feature flags**, **Fields**, **Flows**,
+**Pages**, and **Workspace settings**. **Flows** holds **Step Layouts**, and **Pages**
+holds **Page Layouts** and **Sections**.
+
+A step, an action, and a condition have no sidebar entry. You reach all three from the
+library in the flow builder, because each one belongs to a flow.
+
+## API and MCP
+
+You can also do this task with the [Management API](/reference/management-api/schema), the [MCP server](/reference/tools/mcp), or code that you send to [`picked_run`](/reference/tools/mcp-run).
+
+| Task | GraphQL | MCP tool | Run |
+| --- | --- | --- | --- |
+| List the projects | [`projects`](/reference/management-api/schema/project#projects) | [`list_projects`](/reference/tools/mcp#list_projects) | `tools.list_projects()` |
+| List the workspaces in a project | [`project`](/reference/management-api/schema/project#project) | [`list_workspaces`](/reference/tools/mcp#list_workspaces) | `tools.list_workspaces({ project_id })` |
+
+## Related
+
+- [How to create a project](/how-to/create-a-project)
+- [How to install the SDK](/how-to/install-the-sdk)

package/docs/reference/dashboard/reports.mdx

@@ -0,0 +1,53 @@
+---
+type: reference
+layout: docs
+---
+
+import { Screenshot } from '../../../../components/Screenshot'
+
+# Reports
+
+A report is an embedded Metabase dashboard or question, attached to a Picked project. Use
+a report to see how your visitors move through your experiences and your feature flags.
+
+<Screenshot
+  route="/{orgSlug}/{projectId}/reports"
+  alt="The reports list for a project on the Picked dashboard"
+/>
+
+## Report options
+
+| Option | Description |
+| --- | --- |
+| **Report name** | The name that the reports list shows. |
+| **Description** | An optional note about what the report shows. |
+| **Metabase report type** | `Dashboard` or `Question`. |
+| **Metabase report ID** | The numeric ID from the Metabase URL. |
+| **Display order** | Optional. It controls the position of the report in the list. |
+
+The create form has a link that opens the report, so you can check that the ID is
+correct before you save.
+
+## Permissions
+
+A member needs permission in the organization to create, edit, or delete a report. The
+dashboard hides those actions from a member without it.
+
+## Delete a report
+
+Open the report settings page and click **Delete report**. You cannot reverse a delete.
+
+## API and MCP
+
+You can also do this task with the [Management API](/reference/management-api/schema), the [MCP server](/reference/tools/mcp), or code that you send to [`picked_run`](/reference/tools/mcp-run).
+
+| Task | GraphQL | MCP tool | Run |
+| --- | --- | --- | --- |
+| Create a report | [`reportCreate`](/reference/management-api/schema/report#reportcreate) | None | None |
+| Edit a report | [`reportUpdate`](/reference/management-api/schema/report#reportupdate) | None | None |
+| Delete a report | [`reportDelete`](/reference/management-api/schema/report#reportdelete) | None | None |
+
+## Related
+
+- [Analytics queries](/how-to/analytics) — query the same data with SQL.
+- [About PickedQL](/explanation/how-pickedql-works) — what the query layer expands.

package/docs/reference/dashboard/steps.mdx

@@ -0,0 +1,47 @@
+---
+type: reference
+layout: docs
+---
+
+import { Screenshot } from '../../../../components/Screenshot'
+
+# Steps
+
+A step is one screen or card that a visitor sees during an experience. Each step
+holds one or more fields. A field collects data or shows content.
+
+<Screenshot
+  route="/{orgSlug}/{projectId}/workspaces/{workspaceId}/steps"
+  alt="The steps list for a workspace on the Picked dashboard"
+/>
+
+## Set up a step
+
+**Name**: The name that the dashboard shows. Your visitors do not see it.
+
+**Key**: The identifier that your application uses for this step. Your
+visitors do not see it.
+
+**Layout Name**: The name of the component layout that renders this step. Examples
+are `card`, `modal`, and `fullscreen`.
+
+A step can also carry a **Step Layout**. A step layout is a reusable CMS schema, and
+it gives the step a Content Entry. See [Models](/reference/dashboard/content/models).
+
+## API and MCP
+
+You can also do this task with the [Management API](/reference/management-api/schema), the [MCP server](/reference/tools/mcp), or code that you send to [`picked_run`](/reference/tools/mcp-run).
+
+| Task | GraphQL | MCP tool | Run |
+| --- | --- | --- | --- |
+| Create a step | [`stepCreate`](/reference/management-api/schema/step#stepcreate), [`experienceFlowUpdate`](/reference/management-api/schema/experience#experienceflowupdate) | [`update_experience_flow`](/reference/tools/mcp#update_experience_flow) | `tools.update_experience_flow({ experience_id, workspace_id, create_steps, create_nodes })` |
+| Change the layout name of a step | [`stepUpdate`](/reference/management-api/schema/step#stepupdate), [`experienceFlowUpdate`](/reference/management-api/schema/experience#experienceflowupdate) | [`update_experience_flow`](/reference/tools/mcp#update_experience_flow) | `tools.update_experience_flow({ experience_id, workspace_id, update_steps })` |
+| Give a step a Content Entry | [`stepUpdate`](/reference/management-api/schema/step#stepupdate) | [`update_step`](/reference/tools/mcp#update_step) | `tools.update_step({ step_id, layout_content_entry_id })` |
+
+## Related
+
+- [How to add a step, a field, an action, or a condition](/how-to/add-a-step-field-or-action)
+
+- [How to build a flow](/how-to/build-a-flow) — add a step and assign fields to it.
+- [How to reuse a step, a field, or an action](/how-to/reuse-a-step-field-or-action)
+- [Models](/reference/dashboard/content/models) — the step layout schema.

package/docs/reference/dashboard/team-management.md

@@ -0,0 +1,38 @@
+---
+type: reference
+layout: docs
+---
+
+# Team management
+
+Control who can reach your organization and what each person can do. An organization
+is the top-level container for your projects and members.
+
+## Organizations
+
+An organization groups your projects, your members, and your settings. You can belong
+to more than one organization.
+
+To move between organizations, use the organization control at the top of the
+sidebar.
+
+## Roles and permissions
+
+Each member holds a role, and the role controls what the member can reach. A role
+decides whether a member can create a project, edit an experience, control other
+members, or read a report.
+
+Give each member the smallest role that the work of that member needs.
+
+## API and MCP
+
+You can also do this task with the [Management API](/reference/management-api/schema), the [MCP server](/reference/tools/mcp), or code that you send to [`picked_run`](/reference/tools/mcp-run).
+
+| Task | GraphQL | MCP tool | Run |
+| --- | --- | --- | --- |
+| List your organizations | [`organizations`](/reference/management-api/schema/organization#organizations) | [`list_projects`](/reference/tools/mcp#list_projects) | `tools.list_projects()` |
+| Set what a member can do | [`organizationMemberScopesUpdate`](/reference/management-api/schema/organization#organizationmemberscopesupdate) | None | None |
+
+## Related
+
+- [How to invite and remove members](/how-to/invite-and-remove-members)

package/docs/reference/dashboard/workflow-runs.md

@@ -0,0 +1,157 @@
+---
+type: reference
+layout: docs
+---
+
+# Workflow runs
+
+A run is one execution of a published workflow. An event or an action starts it. Picked
+records each run, each node that the run reached, and each lead that the run created or
+changed.
+
+## Where runs show
+
+| Page | Address                             | Content                                                                                      |
+| ---- | ----------------------------------- | -------------------------------------------------------------------------------------------- |
+| Runs | `…/workflows/<workflow>/runs`       | The runs of one workflow, the last run first. Open it with the **Runs** tab of the workflow. |
+| Run  | `…/workflows/<workflow>/runs/<run>` | One run: its records, its tasks, its details and its labels.                                 |
+| Lead | `…/leads/<lead>`                    | The **Workflow runs** card. It lists the runs that created or changed the lead.              |
+
+A run starts only from a published version of the workspace. A change to a workflow that
+you did not publish does not start runs.
+
+## Run status
+
+| Status      | Meaning                                                                                         |
+| ----------- | ----------------------------------------------------------------------------------------------- |
+| `Running`   | The run started and did not stop yet. A **Wait** node can keep a run in this status for days.   |
+| `Completed` | Each node that the run reached did its work, or failed and the run continued on its error exit. |
+| `Failed`    | A node failed. The run stopped at that node. The run shows the reason.                          |
+
+A tool node that fails with an edge on its error exit shows `Failed` on its task. The run
+continues on that edge only when the tool itself fails:
+
+- An External API call fails. Refer to the [External APIs reference](/reference/dashboard/external-apis#errors).
+- Another tool refuses the request. For example, a label value is not a number for a
+  `number` label, or the lead does not exist. The tool also refuses a row with `increment`
+  on a label that is not a `number` label. It refuses some project labels and run labels
+  with a key in the format of a hash. Refer to the [Labels reference](/reference/dashboard/labels).
+
+When the tool refuses the request, the error record has the code `tool_failed` and a
+reason from the tool, for example `value_invalid_for_type`, `label_write_invalid` or
+`owner_not_found`.
+
+Other failures stop the run, with or without an edge on the error exit:
+
+- A required argument has no value.
+- An argument has a value that the tool does not accept:
+  - The value has the wrong type.
+  - The value is not one of the options.
+  - A number is not in the range that the tool accepts.
+  - A label namespace or key does not have the correct format.
+  - Two labels have the same namespace and key.
+  - A list of labels is empty.
+  - A label row has `value` and `increment`, or neither.
+  - A label row has `increment` and `onlyIfAbsent`, or `increment` and a type other than
+    `number`.
+  - An `expiresIn` is not a whole number from 1 to 31536000.
+- A template cannot be rendered.
+- The workflow does not have a scope that the tool needs.
+- The tool gives output that is not valid.
+- An error occurs in Picked.
+
+## Visit cap
+
+A run can go back to a node that it reached before, on a loop. Each time that the run gets
+to a node is a visit. The visit cap is the maximum number of visits to one node in a run.
+
+| Setting       | Value                                                              |
+| ------------- | ------------------------------------------------------------------ |
+| **Visit cap** | A number from 1 to 100. If you do not set it, the visit cap is 10. |
+
+Set the visit cap in the workflow settings. In the definition of the workflow, the field is
+`maxNodeVisits`.
+
+A node can get to the visit cap. If the run then gets to the node again, the run stops.
+The run shows `Failed`, with the code `loop_limit_reached` and the ID of the node. Refer to [How to retry a failed call with a loop](/how-to/retry-a-failed-call-with-a-loop).
+
+When you save a workflow, Picked refuses these loops:
+
+- A loop that does not have a sleep node on it.
+- A loop that does not have a condition node on it.
+- An edge from a node to the same node.
+- A loop with the **Action response** node on it.
+
+## Records
+
+The **Records** card lists each lead that the run created or changed. It shows each lead
+one time.
+
+| Label     | Meaning                                                                                            |
+| --------- | -------------------------------------------------------------------------------------------------- |
+| `Created` | A node of the run created the lead. This label replaces `Updated` for the same lead.               |
+| `Updated` | A node of the run changed the lead or its labels.                                                  |
+| `Deleted` | A node of the run deleted the lead. This label replaces `Created` and `Updated` for the same lead. |
+
+When a lead does not exist now, the card shows `Deleted lead` and the ID of the lead,
+with no link.
+
+## Tasks
+
+A task is one visit to a node. The **Tasks** card lists the tasks in the order that the
+run reached them. When the run visits a node more than one time, each visit is a task.
+A task of a later visit shows the number of the visit, such as `visit 2`.
+
+The card shows 50 tasks. **Show more tasks** adds the next 50.
+
+| Column     | Value                                                                                                                                                             |
+| ---------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------- |
+| Name       | The tool of a tool node, or the kind of node: **Start**, **Condition** or **Wait**.                                                                               |
+| Node       | The ID of the node. It links to the node in the editor while the workflow has the node.                                                                           |
+| Status     | `Running`, `Succeeded` or `Failed`.                                                                                                                               |
+| Time       | The time from the start of the task to its end.                                                                                                                   |
+| Operations | The leads and the labels that the task created, changed or removed, the first operation first. A label shows its namespace and its key. It never shows its value. |
+
+A condition task does not show which branch it chose.
+
+## Labels
+
+The **Labels** card shows the labels of the run. A user with `ORGANIZATION_PROJECTS_UPDATE` can
+change, add and remove them. Refer to [Labels in the dashboard](/reference/dashboard/labels#labels-in-the-dashboard).
+
+## Leads and runs
+
+A lead lists a run when a node of that run created the lead, set one of its labels, or
+removed one of its labels. A change to the labels of a lead is a change to the lead.
+
+The **Workflow runs** card shows 20 runs. **Show older runs** adds the next 20.
+
+## Access
+
+| Surface                          | Scopes                                                                                                                                                                        |
+| -------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
+| Runs, Run                        | `ORGANIZATION_WORKSPACE_READ` and `ORGANIZATION_PROJECTS_READ`                                                                                                                |
+| The labels of a run              | `ORGANIZATION_PROJECTS_READ` to read them, and `ORGANIZATION_PROJECTS_UPDATE` to change them                                                                                  |
+| The **Runs** tab of the workflow | `ORGANIZATION_WORKSPACE_READ` and `ORGANIZATION_PROJECTS_READ`. Without these scopes, the workflow does not show the **Runs** tab.                                            |
+| Workflow runs card               | `ORGANIZATION_LEADS_READ` for the lead, and `ORGANIZATION_WORKSPACE_READ` and `ORGANIZATION_PROJECTS_READ` for its runs. Without the last two scopes, the card does not show. |
+
+## Deleted workflows
+
+A run stays after you delete its workflow. The run page and the lead card show it as
+**Deleted workflow**. The **Runs** page of a deleted workflow does not exist, and the run
+page has no link to the editor.
+
+## API and MCP
+
+You can also do this task with the [Management API](/reference/management-api/schema), the [MCP server](/reference/tools/mcp), or code that you send to [`picked_run`](/reference/tools/mcp-run).
+
+| Task | GraphQL | MCP tool | Run |
+| --- | --- | --- | --- |
+| List the runs of a workflow | [`workflow`](/reference/management-api/schema/workflow#workflow) | [`list_workflow_runs`](/reference/tools/mcp#list_workflow_runs) | `tools.list_workflow_runs({ workflow_id })` |
+| Read a run | [`workflowRun`](/reference/management-api/schema/workflow-run#workflowrun) | [`get_workflow_run`](/reference/tools/mcp#get_workflow_run) | `tools.get_workflow_run({ run_id })` |
+| List the tasks of a run | [`workflowRun`](/reference/management-api/schema/workflow-run#workflowrun) | [`list_workflow_run_tasks`](/reference/tools/mcp#list_workflow_run_tasks) | `tools.list_workflow_run_tasks({ run_id })` |
+| List the runs of a lead | [`lead`](/reference/management-api/schema/lead#lead) | None | None |
+
+The `workflow` query gives the runs in its `runs` field. The `workflowRun` query gives
+the tasks in its `tasks` field. The `lead` query gives the runs in its `workflowRuns`
+field.