All skills
simota avatar

/cast

@35ffd55
by shingo imotasimota/agent-skills85 stars
15

Casting personas: rapid generation from diverse inputs, registry-based persistence and lifecycle, data-driven evolution, inter-agent sync. Not for UI walkthroughs (Echo) or user research (Field).

Use this Skill: https://skilld.dev/gh/simota/agent-skills/cast

This session only. Nothing lands on disk.

referencearchetype-mapping.md

≈3k tokens on demand. Your agent reads this file only when SKILL.md points to it.

Archetype Mapping Reference

Purpose: Tag personas with brand and behavioral archetypes so downstream agents (Echo, Spark, Voice, Growth) inherit consistent narrative DNA. Cover Jung 12 brand archetypes, Jobs-To-Be-Done archetype model, persona-archetype mapping, and brand consistency validation.

Scope Boundary

  • cast archetype: Brand + JTBD archetype tagging on personas (this document).
  • cast generate (elsewhere): Persona generation from sources.
  • Saga (elsewhere): Customer story narratives (uses archetypes downstream).
  • Vision (elsewhere): Brand identity direction.
  • Field (elsewhere): Foundational user research.

Two Archetype Layers

Layer Purpose Source
Brand Archetype Persona's emotional / aspirational identity (who they want to be) Carl Jung / Margaret Mark + Carol Pearson
JTBD Archetype Persona's job-when-they-hire-the-product (functional / emotional / social) Clayton Christensen / Bob Moesta

A persona ideally maps to both: Brand archetype for tone of voice, JTBD archetype for feature priority.

The 12 Jung Brand Archetypes

# Archetype Core desire Voice Example brands
1 Innocent Safety, simplicity Optimistic, sincere Coca-Cola, Dove
2 Magi Truth, understanding Authoritative, thoughtful Google, BBC
3 Explorer Freedom, discovery Adventurous, restless Patagonia, Jeep
4 Outlaw Liberation, disruption Rebellious, blunt Harley-Davidson, Virgin
5 Magician Transformation, vision Mystical, inspiring Apple, Tesla, Disney
6 Hero Mastery, achievement Bold, confident Nike, BMW
7 Lover Intimacy, beauty Sensual, warm Chanel, Häagen-Dazs
8 Jester Joy, playfulness Funny, irreverent Old Spice, M&M's
9 Everyman Belonging, fairness Friendly, down-to-earth IKEA, Levi's
10 Caregiver Service, compassion Nurturing, supportive Johnson & Johnson, Volvo
11 Ruler Control, status Refined, commanding Rolex, Mercedes
12 Creator Innovation, expression Visionary, expressive Adobe, LEGO

Three motivational families:

  • Order: Caregiver, Ruler, Creator (provide structure)
  • Belonging: Lover, Jester, Everyman (foster connection)
  • Change: Hero, Outlaw, Magician (mastery + transformation)
  • Meaning: Innocent, Magi, Explorer (independence + truth)

Change-agent personas (Echo[demand] Disposition axis). Echo[demand]'s transformation-voicing personas — Entrepreneur, Revolutionary, Maverick/Heretic, Early-adopter visionary (echo/reference/demand-persona-embodiment.md § Challenger Archetypes) — are the registry home for bold, anti-incremental demand sources. Tag them with the Change / Explorer family: Entrepreneur → Creator/Magician/Hero · Revolutionary → Outlaw/Magician · Maverick → Outlaw/Explorer · Visionary → Explorer/Magician. They still require ≥3 evidence quotes (or synthetic: true + confidence ≤ 0.50 when proto-generated) and a Change/Explorer-compatible JTBD (typically emotional/social) — do not assign Outlaw + Functional-only.

JTBD Archetypes

            FUNCTIONAL
                |
                |  "Help me get this job done"
                |
EMOTIONAL ──────┼────── SOCIAL
                |
"Help me feel"  |  "Help me be seen as..."

Per Christensen, every "hire" of a product satisfies a mix of:

  • Functional: solve a tangible task ("file taxes correctly")
  • Emotional: feel a way ("feel competent, not anxious")
  • Social: be seen a way ("be seen as responsible")

JTBD archetype categorizes the dominant dimension. A persona may have multiple jobs; identify the primary one for product positioning.

Persona ↔ Archetype Mapping

persona:
  id: P-042
  name: Aoi Mori
  brand_archetype:
    primary: Magi
    secondary: Caregiver
    rationale: |
      Wants accurate information first, then to share it with others.
      Quotes: "I read 3 reviews before any purchase" / "I always tell my colleagues."
  jtbd_archetype:
    job_dimension: emotional
    job_statement: "When I'm evaluating a complex tool, I want to feel confident that I haven't missed anything, so I can recommend it without reservation."
    functional_aspect: thorough comparison
    emotional_aspect: confidence, low regret
    social_aspect: trusted advisor

Selection Heuristics

RAW PERSONA EVIDENCE  →  ARCHETYPE TAG
──────────────────────────────────────
"I research extensively"           → Magi
"I want to feel free / explore"    → Explorer
"I need to look successful"        → Ruler / Hero (depending on tone)
"I want simple, no surprises"      → Innocent / Everyman
"I love clever / playful brands"   → Jester / Creator
"I take care of others"            → Caregiver
"I disrupt the status quo"         → Outlaw
"I want transformation"            → Magician

Use evidence quotes from interviews / reviews / support tickets — never assign by intuition alone.

Brand-Archetype Consistency Check

For multi-persona sets, validate that the brand the personas serve has a coherent archetype mix:

Brand archetype Dominant persona archetypes
Apple (Magician) Magician + Creator + Explorer
Patagonia (Explorer) Explorer + Caregiver
Stripe (Magi) Magi + Creator

If the persona set has ≥3 dominant archetypes that don't ladder up to a coherent brand, the segmentation is wrong.

Workflow

EVIDENCE     →  collect quotes / behaviors per persona
             →  tag each behavior with motivational family

JUNG-MAP     →  primary brand archetype (single)
             →  secondary archetype (modifier)
             →  rationale with ≥3 evidence citations

JTBD-MAP     →  identify job statement (When __, I want __, so I can __)
             →  classify dominant dimension (F/E/S)
             →  enumerate F + E + S aspects

CONSISTENCY  →  brand-level archetype coherence
             →  flag personas that don't ladder up
             →  brand archetype = mode of persona archetypes

VALIDATE     →  evidence-cited quotes per archetype
             →  no archetype assigned without ≥3 quotes
             →  archetype + JTBD compatibility (e.g., Magi+Functional ✓; Outlaw+Innocent ✗)

DELIVER      →  registry update with archetype + jtbd fields
             →  tone-of-voice guide per archetype
             →  feature-priority hint from JTBD dimension

HANDOFF      →  Saga: narrative voice
             →  Spark: feature priority via JTBD
             →  Voice: tone-of-voice copy
             →  Echo: archetype-aligned cognitive walkthrough
             →  Growth: emotional retention triggers

Output Template

## Archetype Mapping: [Persona Set]

### Persona Archetype Table
| ID | Name | Brand Primary | Brand Secondary | JTBD Dim | Job Statement (compact) |
|----|------|---------------|-----------------|----------|--------------------------|
| P-001 | Aoi | Magi | Caregiver | Emotional | feel confident before recommending |
| P-002 | Ren | Hero | Ruler | Social | be seen as the high-performer |
| P-003 | Hana | Caregiver | Innocent | Functional | protect family from financial risk |

### Evidence per Persona
- **P-001 / Magi**:
  - "I always read 3 reviews before buying" (interview T-12)
  - "I keep a Notion of every decision rationale" (behavior log)
  - "I recommend tools to my team on Slack" (Slack mining)
- **P-002 / Hero**: ...

### JTBD Statements
- P-001: "When evaluating a complex tool, I want to feel confident I haven't missed anything, so I can recommend it without reservation."
- P-002: ...

### Brand-Level Coherence
- Modal archetype across personas: [Magi]
- Implied brand archetype: [Magi] — matches stated brand: [yes / no]
- Outliers: [persona IDs that don't ladder up + recommendation]

### Tone-of-Voice Hints
- Magi: authoritative, thoughtful, evidence-led — avoid casual or playful
- Hero: confident, achievement-framed — quantify impact

### JTBD-Driven Feature Priority
- Functional dim → tools, data, integrations
- Emotional dim → reassurance UX, tooltips, undo, confirmation
- Social dim → sharing, badges, public profiles

### Handoffs
- Saga: narrative voice + character sheet
- Spark: feature priority by JTBD dimension
- Voice: per-archetype copy guide
- Echo: walkthrough with archetype-consistent reactions
- Growth: emotional retention triggers per archetype

Anti-Patterns

Anti-pattern Fix
Picking archetype by gut without evidence Require ≥3 evidence quotes per assignment
Multiple primary archetypes per persona One primary + one optional secondary (modifier)
Brand archetype != mode of persona archetypes Re-segment; personas serving different brands shouldn't share a set
JTBD with all three F/E/S equally weighted Pick the dominant dimension; secondary aspects are context
Job statement missing one of the When/Want/So-that clauses Use full template; partial JTBDs are aspirational, not actionable
Stereotyping (Magi = nerdy male, Lover = young female) Bias-audit: representation matrix; archetypes are universal
Archetype + JTBD incompatibility (Outlaw + Functional only) Outlaw is emotional/social; check archetype × JTBD compatibility
Tagging only personas, never the brand Brand archetype enables consistency at portfolio level
Reapplying archetype as a persona name (Persona "The Magi") Archetype is a tag; persona has a real-life-resembling name
Static archetype with no review cadence Re-validate during evolve; market shifts drift archetypes
Using archetypes as creative shortcut without behavior Archetype must trace to evidence; otherwise persona drifts
Forcing Jung 12 onto B2B-only personas Adapt for B2B: Magi / Ruler / Caregiver / Hero are typically dominant

Deliverable Contract

When archetype completes, emit:

  • Per-persona archetype (primary + secondary, JTBD dimension).
  • Evidence citations (≥3 quotes per archetype).
  • Job statement (When / Want / So-that template).
  • Brand-level coherence check.
  • Tone-of-voice hints per archetype.
  • JTBD-driven feature priority hint.
  • Handoffs: Saga, Spark, Voice, Echo, Growth.

References

  • The Hero and the Outlaw — Margaret Mark + Carol Pearson (12 archetype framework, 2001)
  • Carl Jung — Archetypes and the Collective Unconscious (1959)
  • Competing Against Luck — Clayton Christensen + Karen Dillon + David Duncan + Taddy Hall (2016) — JTBD framework
  • Jobs to be Done — Tony Ulwick (Outcome-Driven Innovation)
  • Demand-Side Sales 101 — Bob Moesta + Greg Engle (JTBD interview techniques)
  • JTBD Pyramid™ — AIM Institute (2025) — addresses five common JTBD practitioner challenges
  • Building a StoryBrand — Donald Miller (brand archetypes for narrative)
  • Designing Brand Identity — Alina Wheeler
  • About Face — Alan Cooper (persona origins)
  • The Persona Lifecycle — Pruitt + Adlin
  • "Brand Archetypes: A Practical Framework" — Iconic Fox / IDEO white papers
  • Jung typology test resources (note: MBTI continues to face significant academic-validity criticism — use only as a creative prompt, not as evidence)
  • ISO 9241-210 — Human-centered design (persona context)
  • Lean UX — Jeff Gothelf + Josh Seiden (proto-personas)
  • Microsoft Inclusive Design — Persona Spectrum approach (inclusive.microsoft.design, refreshed 2024-2025) as a complementary lens to archetype tagging

Source: SKILL.md on GitHub

1 warning13d5 checks · Risk SAFE
  • Gen Agent Trust Hub13d

    The skill is functionally safe but maintains a surface for indirect prompt injection because it processes external repository files and agent data to generate personas without explicit sanitization. It uses standard data science libraries and well-known text-to-speech tools.

  • Socket13d

    No alerts

  • Snyk13d

    Risk: LOW · No issues

  • Runlayer6mo

    3/12 files flagged

  • ZeroLeaks5mo

    Score: 93/100 · 2 sections analyzed

Signed by skilld at 35ffd55. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 2 days ago.

Activeupdated 2 weeks ago

README badge

README badge for simota/agent-skills/cast