Task
Which decision and which user the agent was created for.
Method / no magic
In short: you pick an agent by task, set the boundaries and materials, watch the stages, then review the artifact and its evidence. The model remains part of the system but does not replace the described process, access control and the human decision.
A demonstration is available with synthetic materials and no customer data processing. Full deployment follows an approved estimate.
They let you compare agents by process and output rather than marketing claims.
Which decision and which user the agent was created for.
Which data is required and which is superfluous.
Which stages, tools and control points are used.
The shape of the artifact and the criteria for checking it.
What the result does not prove and where an expert is needed.
Which sources, actions and retention periods are allowed.
Cloud, team, or local deployment defined for the task.
Which tests and traces confirm the declared behavior.
No stage overrides the user's professional judgment.

Compare the passport and the limitations.
Set the goal, data, access and output format.
Follow the stages and answer the questions.
Match the conclusions against the evidence.
Accept, revise, repeat or reject.
A convincing tone does not count as evidence.
Every key conclusion can be traced back to a source or an assumption.
Facts, interpretations and recommendations are not mixed.
Gaps, exceptions and contradictions are visible.
The recommendation is turned into an observable test.
It is known which data and conditions produced the result.
It is clear who makes the next decision.
First step
Pick one of the twelve agents and subagents, or fill in the questionnaire for custom development.