When to use
Handles text-to-image only. Input: text prompt; does not accept images/Element/image asset references.
When the prompt is not yet finalized, hand it to dreamina-prompt-text2image; do not rewrite the final prompt without the user's request.
Sunset notice: this skill replaces the corresponding generation path of the old dreamina CLI.
Operation Contract
| Item | Standard |
|---|---|
| Input and Prerequisites | User-confirmed prompt, current model spec, projectId, and the references required by the mode |
| Side Effects | Node writes without --run save a draft; paid execution goes through the public execution operation |
| Output Handoff | The returned nodeId and projectId; after generation, also attach submitId and the validated assets |
| Recovery Rules | Recovery uses the original submitId; only with explicit absent + resubmittable may it be resubmitted for recovery under the original identity |
Workflow
Step 1: Reuse dreamina-canvas-cli 's public invocation contract and the confirmed projectId; when installation or login is missing, hand it to the setup/auth skill by name.
Step 2: Validate the mode parameters from the current model/voice query results, and read Parameter Specification. Do not apply the old CLI's model_version, resolution_type, or poll.
Step 3: Save the draft and record the returned nodeId; when modifying an existing node, first read the current state and fully replace the generation parameters and references, only metadata is a sparse update.
dreamina-canvas --format json node create image --project-id <projectId> \
--prompt "<prompt>" --mode t2i --model <model> --ratio <ratio> --resolution <resolution> --count <count>
Step 4: When generation is needed, hand it to dreamina-canvas-cli 's execution operation: node quote → node confirm → node run. Record a stable submitId before submission; the task skill does not approve budgets on its own.
Step 5: Hand it to the same public skill for query recovery and download validation, and verify the actual artifacts of this mode. On timeout, do not resubmit with a different submitId.
Rules
- Placeholder parameters must be replaced with values returned by the current CLI or confirmed by the user; draft commands do not add --run.
- Do not mix required fields from other modes; the batch limit comes from the model capability query.
- Never persist tokens, cookies or signed URLs; recovery retains only non-sensitive identity.
- When a companion skill is missing, install:
npx skills add full-aigc-skills/dreamina-skills --skill <skill-name>。
Progressive disclosure (read on demand)
- When per-flag required/value/conflict rules are needed, read Parameter Specification。
- When selecting a model, ratio, resolution, or duration from the current discovery results, read Model or Voice Selection。
- When a paragraph-style workflow such as batch, async, multi-project, or iteration is needed, read Common Scenario Combinations。
- When encountering timeout, partial success, or recovery scenarios, read Exception Handling。
- When doing a pre-delivery self-check, read Acceptance Checklist。
- For first run, see Success Scenario; for batch/async/multi-project/privilege-escalation-denial/failure-recovery examples, see the
examples/directory.
Validation checklist
- input is consistent with t2i mode, with parameters undergoing real-time discovery.
- The returned projectId, nodeId, and submitId are handed off correctly; acceptance must not be treated as completion.
- Output the final state as successful; after download, verify the media type, byte count, and SHA-256.
Gotchas
- Node generation editing is a full replace and must not lose the original references.
- Login status, billing permissions, and successful generation are different checks.
- Generate only the quantity requested by the user, and do not automatically add attempts.
Not applicable and boundaries
Not applicable to editing with image references; use image2image instead. When only a prompt is requested, hand it off to dreamina-prompt-text2image; when a submitId already exists, follow the public recovery process and do not rebuild the text-to-image task.
Read only the account and asset information necessary for the current operation; do not store, log, or output access tokens, cookies, or temporary signed links.