agentleFS
coval-aicoval-ai/coval-external-skills/skills/personas/design-persona/SKILL.md

design-persona

Skill3 starsChanged 19 days agoMITView on GitHub

Design and create a simulation persona for testing an AI agent. Guides through use case selection, voice and language configuration, behavior prompt crafting, and interruption calibration. Use when user says "create a persona", "design a persona", "set up a test persona", "configure simulation persona", or "build a caller profile".

The file

Show its rules and text

We split this skill into the rules it gives an agent

16 rules. About 0 are common, 14 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈14

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    If $ARGUMENTS contains a use case (e.g.

    line 11
  2. Distinctive:

    If not authenticated, guide the user:

    line 21
  3. Distinctive:

    If the user doesn't have a Coval account, direct them to https://coval.dev to sign up.

    line 27
  4. Distinctive:

    Run these in parallel:

    line 31
  5. Distinctive:

    No personas → proceed to Phase 1

    line 40
  6. Very specific:

    If $ARGUMENTS matches an existing agent name, pre-select that agent for Phase 1

    line 41
  7. Distinctive:

    Ask: "Which agent will this persona test?"

    line 45
  8. Distinctive:

    If agents exist, present them as a numbered list to pick from

    line 47
  9. Distinctive:

    If no agents exist, say: "No agents found.

    line 48
  10. Distinctive:

    If the user selects an agent, fetch its details:

    line 50
  11. Distinctive:

    Ask: "What does your agent do?"

    line 67
  12. Distinctive:

    Ask: "Use this as a starting point?

    line 95
  13. Distinctive:

    Ask: "What interruption rate?

    line 157
  14. Distinctive:

    Start from the template's description as a base prompt.

    line 163
  15. Distinctive:

    Ask: "What specific behavior should this persona exhibit?

    line 167
  16. Distinctive:

    Ask: "Create this persona?

    line 196

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Design Persona

Guide the user through designing and creating a simulation persona using the coval CLI. Follow the phases below in order, asking questions at each step.

If $ARGUMENTS contains a use case (e.g. "insurance_claims", "customer_support") or agent name, use it to skip or pre-fill relevant questions.

Phase 0: Preflight + Inventory

Step 1: Check authentication

coval whoami

If not authenticated, guide the user:

coval login

This prompts for an API key. Get one at https://app.coval.dev/settings (Organization > Manage > API Keys).

If the user doesn't have a Coval account, direct them to https://coval.dev to sign up.

Step 2: Inventory existing resources

Run these in parallel:

coval personas list --format json
coval agents list --format json

Decision matrix:

  • Has personas → present them as a numbered list: "You already have these personas. Want to reuse one, duplicate and modify, or create new?"
  • No personas → proceed to Phase 1
  • If $ARGUMENTS matches an existing agent name, pre-select that agent for Phase 1

Phase 1: Agent Context

Ask: "Which agent will this persona test?"

  • If agents exist, present them as a numbered list to pick from
  • If no agents exist, say: "No agents found. You can still create a persona — just tell me what type of agent it will test (voice, outbound-voice, chat, sms, websocket)."

If the user selects an agent, fetch its details:

coval agents get <agent_id> --format json

Extract from the response:

  • Agent type (model_type) — determines whether voice settings are relevant and conversation initiator direction
  • Agent name — for context in prompts
  • Agent prompt (if available) — helps craft a better persona

Key type mappings:

  • MODEL_TYPE_VOICE / MODEL_TYPE_OUTBOUND_VOICE → voice settings matter, conversation initiator depends on direction
  • MODEL_TYPE_CHAT / MODEL_TYPE_WEBSOCKET / MODEL_TYPE_API / MODEL_TYPE_ENDPOINT → voice settings are defaults only (won't affect simulation)

Phase 2: Use Case Detection

Ask: "What does your agent do?"

Present the options:

  • customer_support — Customer Support
  • scheduling_booking — Scheduling & Booking
  • sales — Sales
  • insurance_claims — Insurance Claims
  • healthcare_intake — Healthcare Intake
  • restaurant_orders — Restaurant Orders
  • debt_collection — Debt Collection
  • it_helpdesk — IT Helpdesk
  • other — Other (describe it)

Load references/persona-templates.md and select the template matching the chosen use case.

Present the template as a starting point:

Here's a starting persona for <use case>:

  Name:       <template name>
  Voice:      <voice>
  Language:   en-US
  Background: <background>
  Wait:       <wait seconds>s
  Prompt:     "<template description>"

Ask: "Use this as a starting point? (yes / customize name / start from scratch)"

Phase 3: Voice + Language

Load references/voice-options.md for available options.

Ask these questions:

  1. "What language should the persona speak?"
    • en-US — English (US)
    • es-ES — Spanish (Spain)
    • fr-FR — French (France)
    • de-DE — German
    • pt-BR — Portuguese (Brazil)
    • ja-JP — Japanese
  2. "Voice preference?"
    • Female: aria (clear, professional)
    • Male: callum (clear, professional)

For non-voice agents (chat, websocket, API), explain: "Since your agent is text-based, voice and language are stored as defaults but won't affect the simulation."

Phase 4: Environment + Behavior

Background Sound

Present options with recommendations based on the use case:

ValueDescriptionRecommended For
quietNo background noiseMedical, legal, financial calls
officeOffice ambient noiseCorporate, business, IT support
cafeRestaurant/cafe noiseCasual, restaurant scenarios
airportAirport/travel noiseTravel-related agents

Say: "Based on your <use case> use case, I'd recommend <recommended>. Use that or pick another?"

Wait Seconds

How long the persona waits before speaking after connection:

ValueStyleBest For
0.3Fast responderRestaurant orders, fast-paced interactions
0.5StandardMost inbound call scenarios
1.0Slow / deliberateOutbound calls, debt collection, elderly callers

Say: "The template uses <template wait>s. Keep that or adjust?"

Interruption Rate

How often the persona interrupts the agent mid-sentence:

RateBehaviorBest For
NONENever interruptsCompliance testing, scripted flows
LOWRare interruptionsStandard testing, polite callers
MEDIUMOccasional interruptionsRealistic conversation simulation
HIGHFrequent interruptionsStress testing, impatient callers

Explain: "NONE is best for validating scripted compliance flows. LOW is realistic for most callers. MEDIUM simulates natural conversation. HIGH stress-tests your agent's ability to handle interruptions."

Ask: "What interruption rate? (NONE / LOW / MEDIUM / HIGH)"

Note: Interruption rate is not yet available as a CLI flag. To set it after creation, use the Coval API: PATCH /v1/personas/{persona_id} with {"interruption_rate": "LOW"}.

Phase 5: Prompt Crafting

Start from the template's description as a base prompt.

Explain the key distinction: "The persona prompt defines WHO the caller is and HOW they behave — their personality, speaking style, and emotional state. It does NOT define what they ask about — that's handled by test cases."

Ask: "What specific behavior should this persona exhibit? For example:"

  • "Speaks quickly and gets frustrated if put on hold"
  • "Elderly caller who needs things repeated"
  • "Non-native speaker who sometimes uses wrong words"
  • "Aggressive caller who threatens to cancel"

Help the user refine the simulated_user_prompt by combining:

  1. The template description as a foundation
  2. The user's language choice applied to {language} placeholders
  3. Any custom behavior the user describes

Present the final prompt for confirmation before proceeding.

Phase 6: Create + Confirm

Present a full summary before creating:

Ready to create this persona:

  Name:              <name>
  Voice:             <voice>
  Language:          <language>
  Background:        <background>
  Wait Seconds:      <wait>
  Interruption Rate: <rate>
  Prompt:            "<final prompt>"

Ask: "Create this persona? (yes / edit)"

Create the persona:

coval personas create \
  --name "<name>" \
  --voice "<voice>" \
  --language "<language>" \
  --prompt "<final prompt>" \
  --background "<background>" \
  --wait-seconds <wait> \
  --format json

Capture persona_id from the JSON response.

Confirm success:

Persona created!

  ID:   <persona_id>
  Name: <name>

  View: https://app.coval.dev/personas/<persona_id>

Phase 7: Next Steps

Suggest what the user can do next:

What's next?

  Build test cases for this persona:     /build-test-suite
  Configure evaluation metrics:          /configure-metrics
  Launch a quick evaluation:             /quick-eval

  Or manage this persona later:
  coval personas get <persona_id>
  coval personas update <persona_id> --prompt "new prompt"

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.