"Operate on Transformer Lab Typer CLI, Textual job monitor, CLI
Scanned 9/9/2026
Install to Claude Code
npx -y skills add VectorSpaceLab/AREX-Skill --skill cli-sdk-workflows --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Cli Sdk Workflows?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/vectorspacelab-cli-sdk-workflows)More formats (shields.io, HTML) on the badges page.
---
name: cli-sdk-workflows
description: "Operate on Transformer Lab Typer CLI, Textual job monitor, CLI
config/auth/profiles, asset transfer helpers, Python SDK
facade/resources/storage, and tfl-remote-trap workflows."
disable-model-invocation: true
metadata:
disco-role: operating
license: AGPL 3.0
---
# CLI and SDK Workflows
Use this sub-skill when a task touches Transformer Lab's Python CLI package, the interactive job monitor TUI, CLI authentication/profile/config state, asset upload/download helpers, the Python SDK facade/resource/storage layers, or the remote task wrapper used by launched jobs.
## Route first
- CLI command parsing, output formats, profile/auth/config, task/job command flows, asset helpers, and Textual monitor behavior: read [CLI reference](references/cli-reference.md).
- SDK task-script APIs, resource classes, storage/organization context, artifact/model/dataset/checkpoint helpers, and `tfl-remote-trap`: read [SDK reference](references/sdk-reference.md).
- Common failures and debugging playbooks: read [Troubleshooting](references/troubleshooting.md).
- Backend API route implementation and service internals belong to `../backend-api-services/SKILL.md`.
- Job lifecycle, queue/provider semantics, remote launch contracts, and compute-provider behavior belong to `../task-execution-compute/SKILL.md`.
- React browser UI behavior belongs to `../frontend-web-app/SKILL.md`.
## Operating rules
1. Keep CLI changes Typer-first unless the workflow truly needs a human interactive TUI. The Textual job monitor is not an automation interface for agents.
2. Preserve root global option ordering: root flags such as `--format`, `--profile`, and root `--no-interactive` must appear immediately after `lab` and before a command group.
3. Preserve both pretty and machine-readable output for non-interactive CLI commands. JSON mode must not mix Rich spinners, tables, prompts, or update-check text into stdout.
4. Keep the CLI independent from the SDK. The CLI talks to the API through its HTTP utility layer; task code running inside jobs imports the SDK.
5. For SDK task examples and generated task templates, call `lab.init()` before logging, progress, or save helpers, and pass final scores as dictionaries, not scalars.
6. Do not solve provider orchestration or backend API behavior in this sub-skill. Route across sibling skills once a question moves below the CLI/SDK facade boundary.
## Fast validation anchors
- CLI package version: `transformerlab-cli` `0.0.68`.
- SDK package version: `transformerlab` `0.1.46`.
- Verified SDK signatures: `Lab.init(self, experiment_id: str | None = None, config: Optional[Dict[str, Any]] = None)` and `Lab.finish(self, message: str = "Job completed successfully", score: Optional[Dict[str, Any]] = None, additional_output_path: Optional[str] = None, plot_data_path: Optional[str] = None)`.
- CLI tests use `typer.testing.CliRunner` with mocked API calls; SDK tests use pytest around isolated workspace/storage state.
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!