Skip to main content
Glama

createStep

Adds a step to a task and appends it to the end of that task's step list. A step is one thing a tester reads, answers, or uploads, except TriggeredEmail: that one is invisible to testers and sends an email the moment a tester answers the step immediately before it, so it takes no text and is rejected as the first step of a task. Creating always appends, so create steps in the order testers should work through them, or reorder them afterwards with updateStepsOrder. On a draft or scheduled beta project the change is applied immediately. On a running or paused beta project the change is STAGED: testers do not see it until commitTaskChanges publishes all staged changes (revertTaskChanges discards them; getBetaTasks with version 'unpublished' shows the staged state). While staging, these tools accept BOTH the live id and the staged id of a task or step, but always respond with the STAGED row — expect ids you did not pass in, with liveVersionId linking back to the live row. Content of finished or archived beta projects cannot be edited.

Input Schema

NameRequiredDescriptionDefault
textNoThe step's question or title, as the tester sees it (up to 128 characters). Omit it for a TriggeredEmail step, which testers never see; passing one is rejected.
typeYesThe kind of step. MultipleChoices, SingleChoice, Rate, Ranking and TriggeredEmail REQUIRE a content shaped for the type (a choices array, a choices array, a scale object, an items array, an email object). FreeText, Instructions and File take no content and reject one. YesNo takes no authored content either: omit it and the fixed Yes / No pair is set automatically, though echoing that exact pair back is accepted so a step read from getBetaTasks round-trips. TriggeredEmail is not a step testers see or answer: it sends an email the moment a tester answers the step immediately before it, so it takes no "text" and can never be the first step of a task.
taskIdYesID of the task, from getBetaTasks. Never invent ids.
contentNoThe step configuration, shaped by type: for MultipleChoices/SingleChoice an array of at least 2 { choice } options with non-empty texts, optionally plus ONE final { choice: "Other", other: true } option that lets the tester type a free answer (its label is fixed: choice must be the literal text "Other", any other wording is rejected); for Rate an object { scale (number, 3 or more), symbol ("Thumb", "Star" or "Square"), lowerText?, higherText? (labels for the scale ends, up to 20 characters, default empty) }; for Ranking an array of at least 2 items { id (a v4 UUID you generate), text } with unique ids and unique texts; for TriggeredEmail an object { to, subject, message, cc? } where to is { allManagers?: boolean, allTesters?: boolean, tagsIds?: string[], personFieldIds?: string[] } and names at least one audience, subject is plain text of at most 128 characters and takes no variables (nothing replaces them there, so the recipient would read the braces), message is the email body as basic HTML reading at most 4096 characters of text, and cc is a list of Person custom field ids. Any other key of the email object or of to is rejected rather than dropped, so bcc and to.emails are not accepted: nothing would send to them. tagsIds are tester tag ids from listTesterTags, and ONLY the entries it marks assignable: a computed smart tag is rejected because a triggered email cannot be sent to one. personFieldIds and cc are ids of Person-type fields from listCustomFields; a personFieldIds field routes the email to the manager-persons that tester recorded in it, cc puts them in copy. The email goes out the first time a tester answers or skips the step before it (skipping a whole optional task sends nothing), one email per tester, and the audiences reach: allTesters ONLY that tester, never every tester in the beta project; tagsIds that same tester, only if they carry one of the tags (tester tags never reach managers, and beta labels are not tags); allManagers every manager on the beta project. To notify a manager when a tester finishes a task, make this the task's last step and use allManagers or personFieldIds. message may use these tester variables: {{USERNAME}}, {{FULL_NAME}}, {{ADDRESS}}, {{EMAIL}}, {{PHONE}}, {{PROJECT_NAME}}, {{WORKSPACE}}, {{REMAINING_PROJECT_DAYS}}, {{CURRENT_WEEK_DAY}}. The {{MANAGER_*}} ones belong to invitations and are rejected here. On an SDK beta project the audience must be managers only (allManagers), because its testers are not workspace testers and every other audience would reach nobody. REQUIRED for MultipleChoices, SingleChoice, Rate, Ranking and TriggeredEmail; omitting it is rejected. Must be OMITTED for FreeText, Instructions and File; passing one is rejected. Omit it for YesNo too (its answers are fixed to Yes / No, set automatically); passing that exact pair back is accepted, anything else is rejected. Pass the array or object value itself, not a JSON-encoded string. Call getBetaTasks to see the current shape of existing steps.
requiredNoWhether the tester must answer this step. Defaults to true.
productIdYesID of the beta project to query. Get valid IDs from the listProducts tool.
descriptionNoSupporting detail shown under the step (up to 1024 characters of text; basic HTML is allowed).

Schema Changelog

Changes observed during successful MCP inspections.

  1. First observed

TDQS

A4.5/5.0
Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Beyond the annotations, which only declare this as a non-read-only, non-destructive mutation, the description discloses substantial behavior: staged versus live changes, staged-id responses with liveVersionId links, TriggeredEmail sending semantics and timing, audience routing constraints, and edit restrictions on finished or archived projects. This is exactly the kind of operational context annotations do not provide.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness3/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is front-loaded and organized, with the core append behavior and staging model stated early. However, it is very long and repeats TriggeredEmail and content-shape details that are also present in the schema, reducing conciseness for an already complex definition.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the tool's complexity, including staging semantics, TriggeredEmail rules, audience restrictions, and extensive content validation, the description is complete enough for an agent to call it correctly. The absence of an output schema is mitigated by the description's note that responses return staged rows and unexpected ids.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 100%, and the description largely restates the detailed schema guidance for content, type, text, and audience fields rather than adding substantial new parameter meaning. It does add useful cross-references and behavioral context, but the parameter-level semantics are already carried by the schema, so the baseline of 3 applies.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description states a specific verb and resource: it adds a step to a task and appends it to the end of that task's step list. It distinguishes the special TriggeredEmail step type and points to sibling tools such as updateStepsOrder and commitTaskChanges, so an agent can identify the operation without opening the schema.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

It explicitly tells the agent to create steps in tester order because creation always appends, and to use updateStepsOrder to reorder afterwards. It also explains immediate application on draft/scheduled projects versus staged changes on running/paused projects, names commitTaskChanges and revertTaskChanges as the publishing/discarding alternatives, and states that finished or archived projects cannot be edited.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Try in Browser

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources