When to Use Opulent

TLDR: Opulent can handle the majority of engineering tasks, including medium and hard complexity work, and the same approach extends to research, operations, and reporting. The clearer and more specific your instructions, the higher the success rate—especially for complex tasks. For broader guidance on working effectively with autonomous agents, see our Knowledge Work 101 guide.

Best Practices

Scope tasks before implementation:

• Explore your workspace, scope the approach, and let Opulent draft a high-context prompt before a single line of code is written. Ask Opulent is the planned dedicated surface for this; today the same scoping happens inside a session.

  Run multiple sessions in parallel:

• Carve out independent tasks and run them simultaneously. Ask Opulent to delegate to managed agents to launch many parallel child sessions at once, or use the Opulent API for programmatic orchestration.

• Return to draft PRs waiting for review.

  Tag Opulent on Slack or Teams:

• Start sessions directly from conversations about bugs, feature requests, or questions. Opulent responds in-thread with updates.

  Let Opulent close the loop:

• Enable Opulent Review with Auto-Fix so Opulent automatically responds to review comments, fixes flagged bugs, and iterates on CI failures—without you needing to be in the loop. The result: PRs that are ready to merge by the time you look at them.

  Extend Opulent's reach with the Connector Marketplace:

• Connect Opulent to Datadog, Sentry, databases, Figma, Notion, Stripe, and other tools via approved MCP servers and native connectors (Partial today). Opulent can investigate issues, query data, read designs, and more—all within a single session.

  Let Opulent test its own work:

• Opulent has a full desktop environment with a shell, IDE, and browser. It can spin up your app locally, click through the UI, take screenshots, record screen recordings, and QA its own changes before opening a PR.

  Automate recurring tasks with Scheduled Sessions:

• Set up daily or weekly sessions to triage Sentry errors, update dependencies, generate reports, or any other repeatable work.

  Use the Opulent CLI for local work:

• The Opulent CLI is a planned local surface for working with Opulent directly from your terminal—useful for quick fixes, exploration, and tasks that benefit from your local environment context, with a /handoff to move work to a cloud session when needed. An operator CLI exists today, but the user-facing product CLI is a product target.

Evaluating Tasks for Opulent

When deciding if a task suits Opulent, ask yourself:

1. Can I describe clear success criteria? Tasks with test suites, CI checks, or verifiable outcomes yield the best results.
2. Is there enough context? Provide relevant files, patterns, docs, or examples. The more context, the better.
3. Would breaking this down help? For very large projects, split the work into focused sessions that build on each other. You can run them in parallel with managed agents.

As a rule of thumb: if a task would take you three hours or less, Opulent can most likely do it. For longer tasks, break them into smaller sessions.

Pre-Task Checklist

Task Definition and Scope

• Good tasks have a clear start and end, plus explicit success criteria (e.g., passing tests, matching an existing pattern, CI green)
• For complex tasks, scope the work collaboratively before starting a session—investigate the workspace and outline your approach first. Ask Opulent is the planned surface for this.

Available Context

• Are there examples or patterns for Opulent to follow?
• Can you provide prototypes, partial work, or existing patterns from the workspace or docs?
• Are there links, filenames, or design files for Opulent to reference?
• Have you connected relevant tools via the Connector Marketplace (databases, monitoring, design tools)?

Success Validation

• Tasks with test suites, lint checks, or compilation steps yield better results
• Opulent can test its own work by launching your app and verifying behavior in the browser
• Enable Opulent Review to catch bugs before you even look at the PR

Review Effort

• With Auto-Fix enabled, Opulent responds to review comments and CI failures automatically
• Ideally, you just need to see that CI passes and the PR is approved

Task Size

• For large tasks, consider breaking them down into sub-tasks or running them in parallel across managed agents
• Splitting large requests into smaller, manageable chunks helps Opulent stay on track
• Try to keep sessions focused (XS, S, or M as measured by Session Insights)

Post-Task Review

Monitor Session Trajectory

• Use Session Insights to investigate the session timeline and identify actionable feedback for future sessions

• If Opulent repeatedly encounters session usage limits, the task assigned to it might be too complex
• If Opulent is struggling with its dev environment, revisit the Workspace setup

Learning from Opulent's Mistakes

• In your future sessions, provide more context or instructions to help Opulent get past previous obstacles
• Consider adding or approving Knowledge so Opulent remembers things it learned from previous sessions
• Use the improved prompt suggested by Session Insights as a starting point for similar future tasks