Meet the agents

Open a New Session. Bright creates a new tab — useful because you’ll often work on multiple threads at once.

Ask what’s available

What agents / subagents are available in Brightstack? And explain what each agent does.

Agent	Best for
Manager	Routes work to the right specialist and synthesizes the final answer
Product Manager	Specs, issues, acceptance criteria, scoping, prioritization
Engineer	Code, architecture, repo work, sandbox, tests, debugging
Researcher	Deep multi-source comparisons and recommendations
Explorer	Fast, bounded lookups (one web/workspace search at a time)
Designer	UI/visual direction and design review (text guidance, not code)
Assistant	Small misc tasks: docs cleanup, labels, formatting, config tweaks

These agents implement Bright’s loop process:

Define the goal

Take your goal as input, or help clarify it.

Gather context

Inspect the site, repo, docs, or external sources as needed.

Take action

Research, write code, update docs, create issues — depending on mode and task.

Verify

Check results against goals and constraints.

Repeat

Loop until the conditions are met.

You’ll also be able to create and customize agents later. The set above is the out-of-the-box starting team.

See the loop in action with Ask mode research.