Skip to content

datarobot-oss/datarobot-agent-skills / skills/datarobot-agent-assist/agent-assist-build

datarobot-agent-assist-build

Flagged

Use when the user wants to design, build, code, or deploy an AI agent on DataRobot; mentions agent_spec.yaml, dress rehearsal, the DataRobot agent template, LangGraph, CrewAI, LlamaIndex, NAT, Base agents, MCP servers, backend APIs, custom frontends, or the DataRobot CLI.

references/helper-scripts.md

View on GitHub
## Helper Scripts

Scripts live in `<skill_scripts_dir>` (`scripts/` next to `SKILL.md`). Resolve the path once per session — see [Script Path Resolution](../SKILL.md#script-path-resolution).

The canonical template URL is the `REPO_URL` constant in `clone_template.py` — use it for remote comparison; do not hardcode the URL elsewhere.

### `.env` file

Project `.env` belongs at `<target_dir>/.env` only. Scripts that read or create it require `--target-dir <target_dir>`:

| Script | When |
|--------|------|
| `list_llm_models.py` | Design — model selection |
| `rehearsal.py` | Design — dress rehearsal (`--init` and optional turn override) |
| `setup_template.py` | Pre-coding — template setup (step 9) |

Do not run `dr dotenv setup` manually in cwd when designing in a subdirectory.

### clone_template.py

Clones the DataRobot agent application template repository (URL and tag are defined in the script):

```bash
python <skill_scripts_dir>/clone_template.py \
  --target-dir <target_dir>
```

### list_llm_models.py

Lists the LLMs available on the instance the project's `.env` points at, from two sources: the LLM Gateway catalog (`source: gateway`) and existing DataRobot text-generation deployments (`source: deployed`). Sourced from `dr llm-gateway list`, with a direct REST call as fallback.

```bash
python <skill_scripts_dir>/list_llm_models.py \
  --json \
  --target-dir <target_dir>
```

Pass a `gateway` entry's `llm_default_model` to `setup_template.py --llm-model`. That is the only field that works as a model name: `id` is the catalog's `llmId`, which the gateway rejects, and `api_model` lacks the `datarobot/` prefix that routes the call to DataRobot.

Every `deployed` entry reports the same `llm_default_model` placeholder, so `deployment_id` is what identifies one; `name` is its deployment label.

If the listing looks like it came from the wrong DataRobot instance, compare the `listing requested from` line against the host named in the CLI log lines that follow it. The CLI honors the credentials passed to it only once they verify, and otherwise falls back to its stored profile, so a stale project `.env` yields another instance's catalog rather than an error.

### setup_template.py

Sets up a template repository for initializing a new agent project:

```bash
python <skill_scripts_dir>/setup_template.py \
  --llm-model <model-name> \
  --target-dir <target_dir>
```

For a DataRobot-deployed LLM, pass the spec's `llm_deployment_id` as well:

```bash
python <skill_scripts_dir>/setup_template.py \
  --llm-model <model-name> \
  --llm-deployment-id <deployment-id> \
  --target-dir <target_dir>
```

That writes `LLM_DEPLOYMENT_ID`, `INFRA_ENABLE_LLM=deployed_llm.py`, and `USE_DATAROBOT_LLM_GATEWAY=0` into `.env`, which is what makes the template route to the deployment. Passing the `datarobot-deployed-llm` placeholder as `--llm-model` **without** an id is refused: the template would stay on its gateway configuration and fail much later at `pulumi up` with `Model 'datarobot-deployed-llm' not found in catalog`.

On the deployed path the deployment id is the only thing that selects the model. `dr dotenv setup` rebuilds `.env` from the template's `.datarobot/cli/llm.yml`, whose deployed-LLM group does not include `LLM_DEFAULT_MODEL`, so that key is dropped and the template falls back to its own `datarobot/datarobot-deployed-llm` placeholder. Routing is unaffected; a real model name passed as `--llm-model` alongside an id does not survive.

### select_framework.py

Saves the chosen agentic framework to `.datarobot/answers/agent-agent.yml` (`agent_template_framework`). Preserves all other fields in the file.

```bash
python <skill_scripts_dir>/select_framework.py \
  --framework <value> \
  --target-dir <target_dir>
```

Valid `--framework` values: `langgraph`, `crewai`, `llamaindex`, `nat`, `base`

### check_codespace.py

Used in the Pre-requisite Check (no-op outside a DataRobot Codespace):

```bash
python <skill_scripts_dir>/check_codespace.py
```

### rehearsal.py

Dress rehearsal engine — see [dress-rehearsal.md](dress-rehearsal.md).

```bash
# Initialize (creates session under <target_dir>/.datarobot/rehearsal/<session_id>/)
python <skill_scripts_dir>/rehearsal.py --init \
  --spec <target_dir>/agent_spec.yaml \
  --target-dir <target_dir>

# Turn
python <skill_scripts_dir>/rehearsal.py --session <session_dir> "user message"

# Record a design note (also triggered by NOTE: during rehearsal)
python <skill_scripts_dir>/rehearsal.py --session <session_dir> --note "observation"

# Generate shareable report (default: <target_dir>/rehearsal_report/rehearsal_report.md)
python <skill_scripts_dir>/rehearsal.py --report --session <session_dir>
python <skill_scripts_dir>/rehearsal.py --done --session <session_dir>
python <skill_scripts_dir>/rehearsal.py --session <session_dir> "DONE"
python <skill_scripts_dir>/rehearsal.py --report --session <session_dir> --transcript full
```

Report output:

- Latest: `<target_dir>/rehearsal_report/rehearsal_report.md`
- Archive: `<target_dir>/.datarobot/rehearsal/<session_id>/rehearsal_report.md`

The orchestrating agent appends `## Suggested Changes` (and optionally `## Spec Updates Applied`) after the script writes the base report.
writing-skillsobra/superpowersUse when creating new skills, editing existing skills, or verifying skills work before deploymentStars: 323kUpdated Claude CodeCodexVerifiedInstall
xlsxanthropics/skillsUse this skill any time a spreadsheet file is the primary input or output. This means any task where the user wants to: open, read, edit, or fix an existing .xlsx, .xlsm, .xltx, .csv, or .tsv file (e.g., adding columns, computing formulas, formatting, charting, cleaning messy data); create a new spreadsheet from scratch or from other data sources; or convert between tabular file formats. Trigger especially when the user references a spreadsheet file by name or path — even casually (like "the xlsx in my downloads") — and wants something done to it or produced from it. Also trigger for cleaning or restructuring messy tabular data files (malformed rows, misplaced headers, junk data) into proper spreadsheets. The deliverable must be a spreadsheet file. Do NOT trigger when the primary deliverable is a Word document, HTML report, standalone Python script, database pipeline, or Google Sheets API integration, even if tabular data is involved.Stars: 198kUpdated Claude CodeCodexVerifiedInstall
shipping-and-launchaddyosmani/agent-skillsPrepares production launches. Use when preparing to deploy to production, or when asking what needs to be in place before shipping. Use when you need a pre-launch checklist, when setting up monitoring, when planning a staged rollout, or when you need a rollback strategy.Stars: 119kUpdated Claude CodeCodexVerifiedInstall
childhoodthedotmack/claude-memBabysit a shipped release after merge, publication, or deployment. Verify published packages, production behavior, installed plugins, and late CI or user reports until the observation window is clean; use babysit for PR readiness before release.Stars: 118kUpdated Claude CodeCodexVerifiedInstall