CypherMonkeyAI omni-channel lab

A lifelong-learner lab on AI omni-channel agents, including voice and the contact center. Free to read. Nothing sold.

Prompting and evals · L2 · 15 min

Messages

Learn in five minutes

  • Hosted APIs usually separate system, user, and tool messages. Confirm the current API.
  • The system message is not a secret from a determined user. Do not put a key there.
  • Tool results should be marked as tool results.
  • Order matters because the model reads the sequence.
  • Record which role said each line in the trace.

A trace that flattens every role into one paragraph hides whether the tool or the model spoke. Keep the roles.

This section HTML

Visual

T3-R2 diagram

Diagram SVGDiagram PNGDiagram HTML

Three lines, three roles

  1. Write a system line that sets the job.
  2. Write a user line that asks one question.
  3. Write a tool line that returns a number you computed outside the model.
  4. The reply may quote the tool line and must not invent a second number.

This section HTML

Quiz

Where should a key not go?

A prompt is not a vault.

Why keep roles in the trace?

The trace is the evidence.

What should you confirm about message roles?

APIs change.

This section HTML

Ecosystem

Role
The label that says whether a line is an instruction, a person, or a tool.

Sisters: JSON schema, Rubrics, Regression tests.

The same job-shape-test card works for chat, voice, and a tool result.

The capstone reuses this card.

Config

  • Write the goal in one sentence.
  • Name the input and the allowed output.
  • Save the trace next to the result.

Troubleshooting

  • The answer looks fluent and is wrong. Check whether a source passage actually contains the claim.
  • The loop never stops. Set a step cap and stop when the tool result repeats.

This section HTML

Sources HTML

Sources and go deeper

sources.json