Skip to main content
Agent guide: CLI and SDK instructions →

1. Have your context ready

Provide account identifiers, the question you need answered, required sources, and what counts as current evidence.

2. Ask your agent

Paste this into your agent, then provide your task and example inputs when it asks:
Complete any requested sign-in yourself. Give the agent the missing information before approving execution. Provider-backed work can incur usage.

3. Review the result

The research answers the stated questions with retrievable evidence. Unsupported claims remain open questions, and the saved result preserves account identity. Ask your agent:
Approve a larger job, recurring work, or external changes only after reviewing that decision.

Turn the findings into a workflow

If the research answers your question, keep the cited report. If you need a repeatable workflow, carry the reviewed findings into execution:
If the next step is to define or evaluate an account score, use the scoring guide before implementation.

Check your understanding

Can a missing source response be rewritten as a confident claim that the account has no relevant event?
No. Preserve the source failure or missing evidence. Lack of a retrieved result is not proof of absence.