Skip to content

Sequence Diagram

Jump to bottom
seasonalhawk edited this page Jun 14, 2026 · 1 revision

Sequence Diagram

Normal intake flow

Calling Skill            Elicitation Widget         User
     |                          |                     |
     |-- declares QUESTIONS --> |                     |
     |                          |                     |
     |-- invokes Step 0 ------> |                     |
     |                          |                     |
     |                          |-- choice cards ---> |
     |                          |   (all at once)     |
     |                          |                     |
     |                          | <-- all answers --- |
     |                          |   (one submission)  |
     |                          |                     |
     |                          |-- assembles ----.   |
     |                          |    payload      |   |
     |                          '<----------------'   |
     |                          |                     |
     | <-- payload ------------ |                     |
     |                          |                     |
     |-- proceeds to Step 1 --> |                     |

Step-by-step breakdown

1. Skill declares QUESTIONS

The calling skill defines a QUESTIONS array in its own SKILL.md. This happens at author time, not at runtime. The array is static configuration, not dynamic logic.

2. Skill invokes Step 0

At runtime, Claude reads the calling skill's Step 0 instruction and loads the Elicitation Widget. This is the handoff point. From here the widget is in control.

3. Widget presents all choice cards

The widget calls ask_user_input_v0 exactly once, passing the entire QUESTIONS array. All questions appear together in a single native Claude UI interaction. The user sees every question at once and submits all answers in one action.

This is the critical design constraint: one call, all questions, one submission. The widget never loops. It never calls the tool more than once per intake session.

4. User submits

The user taps their answers across all choice cards and submits once. No back-and-forth. No sequential questioning.

5. Widget assembles payload

After the user submits, the widget collects all answers and assembles a key-value payload:

[Elicitation complete]
Key 1: Value 1 | Key 2: Value 2 | ... | Key N: Value N

This string is not shown to the user. It is passed internally to the calling skill.

6. Calling skill proceeds

The widget hands control back to the calling skill immediately. No pause. No confirmation. The calling skill proceeds to Step 1 with the full payload available as a structured map.

Self-test flow

User                    Elicitation Widget
  |                             |
  |-- "test widget" ----------> |
  |                             |-- announces self-test
  |                             |
  |                             |-- presents 5-question
  |                             |   SELF_TEST_QUESTIONS
  | <-- choice cards ---------- |
  |                             |
  |-- submits all answers ----> |
  |                             |
  | <-- PASSED report --------- |
  |     with assembled payload  |

See Self-Test for full details.

Copyright 2025-2026 Kenneth Benavides. MIT License.

Clone this wiki locally