Rowset Use Cases

Rowset Use Cases

Use this skill to design practical Rowset dataset shapes through the MCP, CLI, or REST interface selected by the user. Treat the patterns below as examples, verify them against the live capability guide and current interface docs, and keep public previews for explicitly requested read-only human sharing.

Before creating or changing datasets:

1. Do not load capabilities or list datasets merely because a session started. For unfamiliar workflows or troubleshooting, load the capability topic index through the selected interface, then request only relevant topics with use cases included. Do not retrieve the full guide unless the workflow needs it.
2. When the user supplies a dataset key or URL, inspect that dataset directly. MCP get_dataset accepts either value. For CLI or REST, extract the dataset key from the URL before using rowset dataset get or /api/datasets/{dataset_key}. Otherwise, search only the relevant datasets or projects with an explicit limit of 3, select one, and inspect its full context before creating anything new.
3. Pick an explicit index column, or let Rowset generate rowset_id when no reliable business key exists.
4. Add dataset instructions, JSON metadata, column descriptions, and choice columns when the agent needs durable operating context.
5. Use relationships when one dataset stores another dataset row's index value.

Personal CRM

Use when the user wants agents to remember people, companies, conversations, and follow-ups.

Starter shape:

• People dataset indexed by email or person_id.
• Companies dataset indexed by company_id.
• Interactions or Messages dataset indexed by message_id.
• Relationship from Messages.person_id to People.person_id.

Useful context:

• People instructions: "Use email as the stable identity. Do not rewrite names from guesses."
• Messages instructions: "Create one row per meaningful interaction. Link to a person when the identity is known."
• Choice columns for relationship stage, follow-up status, or source.

Agent Task Board

Use when the user wants agents to track planned work, blockers, owners, and completion state.

Starter shape:

• Tasks dataset indexed by task_id.
• Choice column status with values such as todo, blocked, doing, done.
• Optional Projects Rowset project for the initiative.

Useful context:

• Dataset instructions define status transitions and when the agent may close a task.
• Metadata can hold status order, default priority, or review rules.
• Use row-by-index updates when task_id is known.

Feedback Triage

Use when the user wants to capture customer feedback, classify it, and track follow-up.

Starter shape:

• Feedback dataset indexed by feedback_id.
• Customers dataset indexed by customer_id or email.
• Relationship from Feedback.customer_id to Customers.

Useful context:

• Choice columns for status, sentiment, source, and priority.
• Metadata for triage policy, escalation rules, or target response time.
• Public preview only if the user wants a read-only summary link.

Content Pipeline

Use when the user wants to manage articles, landing pages, newsletters, or social posts.

Starter shape:

• Content dataset indexed by slug.
• Choice column stage with values such as idea, draft, review, scheduled, published.
• Project metadata linking to docs, repositories, editorial calendars, or source threads.

Useful context:

• Column descriptions for audience, channel, due date, owner, and canonical URL.
• Use exports when another tool needs a full content snapshot.

Product Or Inventory Catalog

Use when the user wants a structured catalog that agents can update and humans can inspect.

Starter shape:

• Products dataset indexed by sku.
• Semantic column types for price, product_url, image_url, and updated_at.
• Optional public preview for read-only sharing.

Useful context:

• Dataset instructions define currency, supplier naming, and retirement rules.
• Choice columns for lifecycle state such as draft, active, retired.

Bug Or QA Tracker

Use when the user wants agents to track issues, repro details, severity, and customer impact.

Starter shape:

• Issues dataset indexed by issue_id.
• Choice columns for status and severity.
• Optional Releases, Customers, or Components datasets with relationships from Issues.

Useful context:

• Instructions define when an issue can move to fixed or verified.
• Metadata can store severity definitions, owner rules, and triage schedule.

Pattern Checklist

For any new use case, choose:

• Dataset names and one stable index per dataset.
• Which values should be choice columns.
• Which column descriptions prevent agent guessing.
• Which relationships should enforce row integrity.
• Which project metadata should store source links or workflow context.
• Which actions are destructive and require explicit user confirmation.