Starter team
Coordinator, Researcher, Reviewer, WriterThat mix gives most new missions enough structure without creating clutter.
Workflow
Operators
Operators are the role layer of opheli.ai. They shape how planning, evidence gathering, review, and packaging happen during a run.
Guide summary
Use this page to understand the surface before you act inside it.
Configure specialist roles such as coordinator, researcher, analyst, reviewer, writer, and coder.
Guide type
WorkflowThis guide reflects the current product workflow and surface ownership.
Sections
8Summary first, then steps, mistakes, and recovery notes.
Related guides
4Written against the current product structure and core execution workflow.
Biggest quality lever
Reviewer roleA reviewer is often what prevents weak claims from becoming a polished artifact.
Related action
Review prompts and model fitOperator quality depends on both instruction quality and provider/model alignment.
Guide section
What it is
Operators are reusable specialist profiles with a role, provider posture, and system prompt that influence how work is executed.
When to use it
Use them whenever you need distinct lanes for coordination, research, analysis, review, writing, or coding.
Where to find it
Find them under Operators in the authenticated product.
What happens next
You use this surface as part of the broader mission -> task -> run -> artifact workflow.
Common mistake
Too many operators
Related action
Open Runs and Mission Replay to see whether the operator team is creating useful handoffs and review pressure.
Guide section
Why it matters
They make execution structured and challengeable instead of relying on one generic assistant voice.
Guide section
How it fits into the workflow
Operators are chosen before or during launch and then participate in run phases.
Guide section
Step-by-step usage
- Create or review the operator role.
- Write a system prompt that reflects the real responsibility of that role.
- Check the provider and model fit for the job.
- Add the operator to the mission or task plan.
- Use Replay and final artifacts to decide whether the role should be refined later.
Guide section
Inputs and outputs
Inputs needed
Role definition
System prompt
Provider/model choice
Mission or task context
Outputs produced
Reusable specialist operators
Clearer execution lanes
Stronger review and packaging posture
Guide section
Common mistakes
- Too many operators
- No reviewer
- Wrong provider/model
- A vague system prompt that describes style but not responsibility
Guide section
Troubleshooting
If runs feel repetitive, generic, or under-challenged, reduce role overlap and strengthen the reviewer lane.
Guide section
Recommended roles
Coordinator
Frames the mission, keeps lanes aligned, and closes the loop at the end.
Researcher
Finds bounded inputs, references, and factual anchors.
Analyst
Turns evidence into structured interpretation and claims.
Reviewer
Challenges weak evidence, risk, and unsupported claims before final packaging.
Writer or Coder
Packages the output into its final form once the work is ready.