Human interaction expert

Overview

Use this skill to turn a product question into an evidence-based HCI method, task model, guideline rationale, or evaluation plan. Frame the user goal, context, and success evidence before selecting an interaction method. Build project-specific artifacts such as a task model, study plan, heuristic ledger, and accessibility acceptance criteria so each recommendation can guide implementation and later evaluation.

Workflow

1. Classify the HCI question and select a method.

• Identify the user group, goal, setting, device, task frequency, consequence of failure, and decision the team needs to make.
• Read references/task_selection.md when the request could use research, task analysis, guidelines, or evaluation.
• Consult references/topic_index.md to connect the question to a method, evidence artifact, guideline source, and oracle.

2. Detect the project shape and establish the evidence baseline.

• Existing project: inventory user journeys, research findings, support issues, analytics, interface states, accessibility checks, and prior study results.
• Greenfield project: write a concise HCI brief with users, goals, context, assumptions, risks, and success measures before interface construction.
• Read references/project_workflow.md for the project-shape path and required HCI artifacts.

3. Model the user's task and mental model.

• Map triggers, goals, actions, decisions, information needs, errors, recovery, handoffs, and completion evidence for each critical task.
• Capture the user's vocabulary, expectations, and visible cues that support recognition, memory, and confidence at each decision point.
• Turn the model into testable task-completion criteria and a study scenario.

4. Select the smallest method that answers the decision.

• Choose interviews or contextual inquiry for unknown context and mental models.
• Choose task analysis for workflow structure and allocation questions.
• Choose cognitive walkthrough or heuristic evaluation for expert inspection.
• Choose usability sessions for observed task completion and comprehension.
• Choose an evaluation study when a hypothesis, comparison, or measured outcome requires controlled or naturalistic evidence.

5. Translate evidence into guidelines and an implementation brief.

• State the applicable cognitive, accessibility, and platform guideline with its user-facing rationale and acceptance criterion.
• Hand the implementation brief to ui-ux-engineer for components, layouts, CSS, interface implementation, visual craft, and state polishing.
• Preserve the task model and rationale while the UI work turns them into a concrete surface.

6. Validate the recommendation with runnable evidence.

• Run guideline-conformance checks, a task-completion walkthrough, and measured accessibility checks; record findings and the next iteration.
• Read references/testing_and_oracles.md for executable checks, measurement methods, and review artifacts.

Implementation defaults

• Start with the local books when present; use references/local_books.md and references/reference_survey.md to select a method backed by the available corpus.
• Use references/topic_index.md as the front door when the corpus is absent, then consult current W3C WCAG and the target platform's human-interface guidelines and run the named oracle.
• Recruit representative participants when a project can conduct a study; otherwise produce an inspection plan and clearly label its evidence scope.
• Write scenarios in the user's language and tie every finding to a task step.
• State measures before collecting evidence: completion, error, time, success, confidence, comprehension, or accessibility conformance.

Quality bar

• Ground recommendations in a user goal, task context, and stated decision.
• Choose a method that can change a product decision within the available time.
• Make each guideline traceable to an observable interaction and acceptance check.
• Separate HCI method ownership from concrete interface implementation ownership.
• Record limitations, participant coverage, and unresolved assumptions clearly.

Output expectations

When using this skill, aim to produce:

• A routed HCI question with the selected method and evidence rationale.
• A task model or study plan with users, scenarios, measures, and completion criteria.
• A guideline ledger that connects cognitive, accessibility, and platform rules to implementation acceptance criteria.
• A runnable validation plan with walkthrough, conformance, and measured accessibility oracles, plus a clear handoff to ui-ux-engineer when needed.