datarobot-oss/datarobot-agent-skills / skills/datarobot-agent-assist/agent-assist-build
datarobot-agent-assist-build
FlaggedUse 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.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 18.9 KB |
| references/agent-spec-examples.md | markdown | 4.8 KB |
| references/agent-spec-schema.md | markdown | 1.8 KB |
| references/dependency-validation.md | markdown | 1005 B |
| references/dr-cli-setup.md | markdown | 872 B |
| references/dress-rehearsal.md | markdown | 7.7 KB |
| references/helper-scripts.md | markdown | 5.1 KB |
| references/llm-selection.md | markdown | 3.5 KB |
| references/pre-coding-checklist.md | markdown | 13.8 KB |
| references/pre-deployment-checklist.md | markdown | 2.2 KB |
| references/resume-design.md | markdown | 7.1 KB |
| references/tool-mapping.md | markdown | 854 B |
| references/workspace-resolution.md | markdown | 7.1 KB |
| scripts/check_codespace.pyexecutable | script | 8.1 KB |
| scripts/clone_template.pyexecutable | script | 7.4 KB |
| scripts/env_utils.pyexecutable | script | 3.9 KB |
| scripts/list_llm_models.pyexecutable | script | 19.6 KB |
| scripts/rehearsal.pyexecutable | script | 44.8 KB |
| scripts/select_framework.pyexecutable | script | 2.4 KB |
| scripts/setup_template.pyexecutable | script | 21.8 KB |
| scripts/write_rehearsal_report.pyexecutable | script | 12.2 KB |
references/workspace-resolution.md
View on GitHub## Workspace Resolution
Run after menu options **1** (Design) or **2** (Code). Sets `<target_dir>` for the session.
For menu option **2**, set `<design_to_code>` = false (cold Code entry). **Post-design → code** skips this document — `<design_to_code>` is set in [Post-design next steps](../SKILL.md#post-design-next-steps) before the pre-coding checklist runs.
**Default subdirectory name:** `new-datarobot-agent`
**Spec filename policy:** Always save new or updated specs as `<target_dir>/agent_spec.yaml`. For every read or existence check, prefer `agent_spec.yaml` and fall back to the legacy `agent_spec.md`; if both exist, use the YAML file. Read a legacy spec as input, but write any changes to `agent_spec.yaml` and do not modify the legacy file.
**After resolution:** all design/code work uses `<target_dir>` — specs at `<target_dir>/agent_spec.yaml`, project `.env` at `<target_dir>/.env`, and helper scripts with `--target-dir <target_dir>` (including `list_llm_models.py`, `rehearsal.py`, and `setup_template.py`). Never create or run `dr dotenv setup` in cwd when `<target_dir>` is a subdirectory.
**On completion** (any path below that sets `<target_dir>`): set `<workspace_resolved>` = true and `<workspace_resolved_target_dir>` = `<target_dir>`.
---
### Path resolution
Use whenever the user must provide a directory containing `agent_spec.yaml` or the legacy `agent_spec.md` (Code entry, or pre-coding Bootstrap step 2 recovery). Referenced from [pre-coding-checklist.md](pre-coding-checklist.md).
1. Ask:
> Which directory contains your agent spec (`agent_spec.yaml` or the legacy `agent_spec.md`)?
2. Do not scan subdirectories automatically. If `./new-datarobot-agent/agent_spec.yaml` exists, you may mention it as a suggestion — do not set `<target_dir>` without the user's answer.
3. Resolve the user's path (relative paths from cwd). If they give a **file path** (e.g. `./my-agent/agent_spec.yaml` or `./my-agent/agent_spec.md`), use its **parent directory** as the project root.
4. The path must be an **existing directory** containing either recognized spec filename. If the directory does not exist or contains neither, ask again.
5. If the user's answer is ambiguous (e.g. "in a subfolder" without a path), ask for the exact directory — do not scan the filesystem to discover specs.
6. Set `<target_dir>` to the resolved directory. Set `<workspace_resolved_target_dir>` = `<target_dir>`. Invalidate `<dependency_check_passed>` (set to false) if it was set.
---
### Decision summary
| Menu | `agent_spec.yaml` or legacy `agent_spec.md` in cwd? | Action |
|------|---------------------------|--------|
| **Design** | Yes | Set `<target_dir>` = cwd. Ask: continue this spec, or new agent in a subdirectory? |
| **Design** | No | Ask: subdirectory (recommended) or current directory → set `<target_dir>` |
| **Code** | Yes | Ask: is cwd the project root? See [Code — spec found in cwd](#code--spec-found-in-cwd) |
| **Code** | No | Ask for project directory → set `<target_dir>`. No subdir scanning. Pre-coding step 2 will not re-ask for the directory if workspace resolution just set it. |
Option **3** (Deploy) skips this document — see [pre-deployment-checklist.md](pre-deployment-checklist.md).
---
### Design — spec found in cwd
1. Set `<target_dir>` to cwd.
2. Resolve the existing spec: use `./agent_spec.yaml` when present, otherwise use `./agent_spec.md`. If both exist, use YAML. Notify the user which file is being continued.
3. Ask:
> Continue editing this spec, or start a new agent in a subdirectory?
- **Continue editing** (or "edit" / "this spec") → read the selected existing file, then read and follow [resume-design.md](resume-design.md). If only the legacy Markdown file exists, preserve its complete YAML content in `<target_dir>/agent_spec.yaml` before making edits or displaying the spec; leave the legacy file unchanged.
- **New agent** (or "new" / "subdirectory") → create/use subdirectory (default `new-datarobot-agent`), set `<target_dir>`, start Design from [Clarification Phase](../SKILL.md#clarification-phase).
### Design — no spec in cwd
Use this path only when neither `./agent_spec.yaml` nor `./agent_spec.md` exists in cwd.
Ask:
> Where would you like to design your agent?
> 1. **Subdirectory** (recommended) — e.g. `./new-datarobot-agent`
> 2. **Current directory**
**Subdirectory chosen:**
- Contains either recognized spec filename → notify, set `<target_dir>`, read the preferred spec, then read and follow [resume-design.md](resume-design.md).
- Contains neither recognized spec filename → set `<target_dir>`, start Design from [Clarification Phase](../SKILL.md#clarification-phase).
- Does not exist → create it, set `<target_dir>`, start Design from [Clarification Phase](../SKILL.md#clarification-phase).
**Current directory chosen:**
- Set `<target_dir>` = cwd.
- If paths other than design-phase artifacts are present, set `<design_messy_cwd>` = true and warn. **Allowed without triggering messy cwd:** `agent_spec.yaml`, legacy `agent_spec.md`, `.env`, `.datarobot/rehearsal/` (dress-rehearsal session cache), and `rehearsal_report/rehearsal_report.md`. Any other file or directory (including `.datarobot/answers/`, source code, or unrelated folders) counts as messy.
- If `<design_messy_cwd>` applies:
> Other files are present. Design can continue here, but coding will require a clean workspace — likely a subdirectory. Files in this directory may be ignored for implementation.
- Start Design from [Clarification Phase](../SKILL.md#clarification-phase).
### Code — spec found in cwd
`agent_spec.yaml` or the legacy `agent_spec.md` exists in cwd, but that alone does not mean cwd is the project root. Prefer the YAML file if both exist. Ask (exact wording):
> An agent spec is in the current directory. Is this your project root?
> 1. **Yes** — use the current directory as the project root
> 2. **No** — the agent is in a different directory
Wait for the user's reply.
- **Choice 1** (or "yes" / "here" / "cwd") → set `<target_dir>` = cwd. Notify: *"Continuing work on `./agent_spec.yaml` in `<target_dir>`."* Proceed to coding.
- **Choice 2** (or "no" / "subdirectory" / "elsewhere" / "different directory") → follow [Path resolution](#path-resolution). Notify: *"Continuing work on `agent_spec.yaml` in `<target_dir>`."* Proceed to coding.
### Code — no spec in cwd
The workspace question below **is** Path resolution step 1. After the user answers, continue with [Path resolution](#path-resolution) steps 2–6 only (do not ask again).
Ask:
> Which project directory contains your `agent_spec.yaml`?
If the user is unsure, remind them the spec may be in a subdirectory (e.g. `./new-datarobot-agent`). If `./new-datarobot-agent/agent_spec.yaml` exists, you may mention it as a suggestion — do not set `<target_dir>` without the user's answer.
Apply Path resolution steps 2–6 to validate and set `<target_dir>`, then set `<workspace_resolved>` and `<workspace_resolved_target_dir>`.
If neither recognized spec filename exists at that path, [pre-coding Bootstrap step 2](pre-coding-checklist.md) offers Design or a single directory correction — it will not repeat the workspace directory question verbatim.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| writing-skillsobra/superpowersUse when creating new skills, editing existing skills, or verifying skills work before deployment | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| 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: 198k | Updated | Claude CodeCodex | Verified | Install |
| 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: 119k | Updated | Claude CodeCodex | Verified | Install |
| 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: 118k | Updated | Claude CodeCodex | Verified | Install |