mental-health-editor

name	mental-health-editor
description	Assessment Editor workflows — edit measure metadata (title, slug, description, version), manage field definitions (add/reorder/remove), and configure scoring rules (thresholds, T-score lookup, reverse scoring).
version	0.1.0
author	hermes-mental-health
license	MIT
metadata	{"hermes":{"tags":["mental-health","editor","authoring","measures"],"category":"mental-health","related_skills":["mental-health-core","mental-health-assessment-review"]}}

Mental Health Assessment Editor

When to use

Use when the practitioner needs to:

• Edit an existing measure's metadata
• Add, remove, or reorder field definitions
• Configure scoring rules (severity thresholds, T-score lookup, reverse items)
• Create a new shared assessment (via Create With AI)

Editor sections

Metadata

• Title — Full DSM-5-TR measure name
• Slug — URL-friendly ID (auto-derived from title)
• Description — Clinical purpose and target population
• Version — Semantic version
• Instructions — Administration guidance shown to patient before form
• resultChart — Which visualization to render for results

Fields

Each field has:

• id — Unique within the measure (e.g., item_1, item_2)
• label — The question/prompt text
• type — scale | text | select | multi_select | boolean
• required — Whether the patient must answer
• options — For select/multi_select: array of {value, label}
• min / max — For scale: the numeric range
• Fields are ordered; reorder support via drag or move-up/move-down

Scoring rules

• calculation — total or average
• maxScale — Highest possible value per item (for reverse scoring)
• severityThresholds — Map of band name → [min, max] range
• reverseScoredItems — Array of field IDs to reverse before totaling
• tScoreLookup — PROMIS raw → T-score mapping table
• requiredFields — Fields that must be answered for the result to be scorable

Create With AI

The Hermes agent generates shared assessment JSON via WebSocket at HERMES_ASSESSMENT_AI_URL. Generated JSON is validated against the Measure Zod schema and persisted to data/shared/assessments/<id>.json.

The agent receives: { assessmentId, prompt, model } and streams JSON output. Never send patient-scoped data to the agent — only the assessment ID + prompt.

Implementation

Component: app/editor/[slug]/ — 4-tab layout (Metadata / Fields / Scoring / Chart)

Editor header

The editor header shows a back link, the assessment title + slug, and the Agent button which links to /agent?editor&slug=${slug} (a dedicated page route, not a modal).

When the agent chat page receives editor + slug query params, the textarea is pre-filled with a context-aware prompt:

Edit the {slug} assessment template using your .hermes agents and mental-health skills. Review the current template configuration and suggest improvements for the metadata, fields, scoring rules, and chart settings.

This pre-fill is implemented in app/agent/_components/agent-chat.tsx using a useState lazy initializer that checks searchParams. From non-editor pages, the textarea starts empty.

Data loading

The server component (page.tsx) loads the full measure from the template JSON via getMeasure(slug) from lib/data/measures.ts (server-only, filesystem). The Measure object is passed as a prop to the client EditorPage component.

Metadata tab

Reads from getMeasureMeta(slug) (client-safe static import from index.json).

Fields tab

Renders each field using FieldRenderer (the same component used in the patient assessment form and result edit mode). Fields are displayed as interactive form elements (scale buttons, selects, etc.) in read-only mode — value={undefined} and onChange={() => {}} to render the widgets without accepting input.

import { FieldRenderer } from "@/app/a/[token]/_components/field-renderer";
// ...
{measure.fields.map((field, idx) => (
  <div key={field.id} className="pb-4 border-b border-border last:border-0 last:pb-0">
    <div className="flex items-start gap-2 mb-1">
      <span className="text-xs text-muted-foreground font-mono shrink-0 mt-0.5">{idx + 1}.</span>
      <FieldRenderer field={field} value={undefined} onChange={() => {}} />
    </div>
  </div>
))}

Scoring tab

Displays full scoring configuration from the template JSON: calculation method, max scale, T-score status, diagnostic threshold, severity thresholds table, reverse-scored items, and required fields.

Chart tab

Renders a Chart Preview using the same custom HTML/CSS severity bar as the result detail page, but with a computed sample score (mid-range, typically Math.floor(fields.length * 1.4)). Shows the full severity bar with colored band segments, a score indicator line, and severity labels below. For t_score_gauge measures, renders a T-score gauge with a sample value of 55.