Create workflow
This skill should be used when the user asks to "create a workflow", "create a getlark test", "add an end-to-end test", "author a larkci workflow", or runs `/getlark:create-workflow`. Converts a natural-language test description (target + ordered steps; target may be a URL, API endpoint, CLI binary, script, or any other software surface) into a `getlark workflows create` invocation with an auto-generated name. Prefer `manage` when the user wants to update or archive an existing workflow, and `invoke-workflow` when they want to run one — this skill only *creates* new workflows.From its SKILL.md
npx -y skills add getlark/skills --skill create-workflowAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
2 things to look at
- no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
- 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its file declares
Copied from the file, not written here
The file declares its own license as MIT. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.
SKILL.md
5.2 KB, as published. Nobody here has run it
create-workflow
Turn a short natural-language test description into a new getlark workflow. The user supplies the description (target + steps). This skill derives a concise workflow name, surfaces optional settings (mode, secret contexts, group), and calls getlark workflows create.
getlark workflows can test any surface — web UIs, HTTP/GraphQL APIs, CLIs, shell scripts, data pipelines, or mixed flows. Do not assume the target is a browser URL.
Inputs
- Description (required) — free-form text containing a target (URL, API base, CLI binary, script path, etc.) and ordered steps. Examples:
- Browser: "Go to https://app.example.com/login, sign in with the
stagingcredentials, click 'New project', confirm the modal, assert dashboard loads." - API: "POST https://api.acme.test/v1/orders with
staging_apicreds and a sample cart payload, assert 201 and that the responseorder.statusispending." - CLI: "Run
mytool import ./fixtures/sample.csv --dry-run, assert exit code 0 and stdout contains3 rows parsed."
- Browser: "Go to https://app.example.com/login, sign in with the
If the user invoked /getlark:create-workflow <text>, treat <text> as the description. Otherwise ask for it.
Procedure
Step 1 — Collect / confirm the description
Ensure the description contains a target (URL, API endpoint, CLI command, script path, etc.) and at least one action step. If either is missing, ask one targeted follow-up. Do not pad thin descriptions with invented steps — confirm with the user first.
Step 2 — Derive a name
Generate a short, Title-Case name (3–6 words) that captures the intent. Heuristic:
- Start from the primary verb + object in the description ("Sign In Flow", "Guest Checkout", "Create New Project").
- Keep under ~50 characters.
- Avoid URL fragments, credentials, or timestamps.
Show the derived name to the user before creating. Offer to accept or override.
Step 3 — Ask about optional fields (only if relevant)
Ask only the questions that matter for this workflow. Skip the rest.
- Mode — default
ai_driven. Only ask if the user mentioned "deterministic", "scripted", "locked", or similar. - Secret contexts — ask if the description references credentials, API tokens, or anything named in quotes that looks like a context (e.g., "the
stagingcredentials"). Offer to list existing contexts viagetlark secret-contexts list. - Group — ask if the user has mentioned organizing tests, or if
getlark workflow-groups listhas existing groups. Otherwise skip.
Use AskUserQuestion for a single batched prompt when more than one optional field is in play.
Step 4 — Invoke the CLI
Build the command with properly quoted arguments:
getlark workflows create \
--name "<derived name>" \
--description "<full description>" \
[--mode deterministic] \
[--secret-contexts name1 name2] \
[--group-id <id>]
Description is passed verbatim — getlark's generation step parses the target and steps server-side. Do not restructure or bullet-ify it.
Note: --secret-contexts takes space-separated values (variadic), not comma-separated.
Step 5 — Report result
The CLI prints the new workflow JSON to stdout. Extract and report:
- Workflow
id status(will typically bepending_generationinitially)- Link:
https://dashboard.getlark.ai/workflows/<id>
Tell the user the workflow is now queued for generation and suggest they run /getlark:invoke-workflow once the workflow status reaches active (or generation_successful). They can check status via getlark workflows get <id>.
Failure modes
--nameor--descriptionmissing: the CLI hard-errors. This skill must always supply both.- Unknown
--group-id: verify withgetlark workflow-groups listbefore passing. - Unknown
--secret-contextsname: verify withgetlark secret-contexts list. The API will reject unknown contexts.
Example
User says: /getlark:create-workflow Go to https://app.example.com/signup, fill the form with a new email, submit, verify confirmation email message appears.
Derive: name = "Signup Flow Confirmation"
Run:
getlark workflows create \
--name "Signup Flow Confirmation" \
--description "Go to https://app.example.com/signup, fill the form with a new email, submit, verify confirmation email message appears."
Report id + dashboard link.