mcpField
P
personas

personas

github.com/scottwaterAgent workflows149 installsUpdated 1w ago
Repository

About personas

personas applies a user-selected persona as a critical lens for evaluating material. It maps names or archetypes to persona files and requires recommendations to be grounded in those files and the supplied evidence.

Key features

  1. 01Match requests to persona files
  2. 02Load persona priorities and style
  3. 03Evaluate material against selected lenses
  4. 04Separate evidence from extrapolation
  5. 05Compare multiple perspectives

Use cases

  1. 01Review a draft through a chosen persona lens
  2. 02Compare feedback from several personas
  3. 03Identify evidence gaps affecting a recommendation

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/scottwater/skills --skill personas

After installing, type/personasin Claude Code to confirm it loaded.

Bundled files

References and scripts installed along with SKILL.md.

agents/
openai.yaml
personas/
bits.md
conversion-copywriter.md
fireball.md
focused-product-manager.md
fried.md
humor-copywriter.md
skeptical-engineer.md
user-advocate.md

Details

From the project README

Persona lenses

Use the requested persona as a critical lens. Ground the response in the persona file rather than inventing opinions or performing a caricature.

Workflow

1. Resolve the requested lens

Match the user's name or archetype to a file under personas/, relative to this skill directory:

FileLens
bits.mdDan Cederholm / SimpleBits
conversion-copywriter.mdJoanna Wiebe / conversion copywriting
fireball.mdJohn Gruber / Daring Fireball
focused-product-manager.mdFocused product manager
fried.mdJason Fried / 37signals
humor-copywriter.mdLianna Patch / humor copywriting
skeptical-engineer.mdSkeptical engineer
user-advocate.mdUser advocate

If the request does not identify a persona, show this list and ask the user to choose. Do not select one for them.

Complete when: every requested lens maps to one persona file.

2. Load the persona

Read each matched file in full. Extract its priorities, recurring questions, positions, and delivery style. Treat named-person files as a bounded synthesis of public ideas, not authority to invent new quotations or claim endorsement.

Complete when: the relevant criteria from every loaded file are represented in working notes.

3. Apply the lens

Evaluate the user's material against those criteria. Tie each material observation to a priority, question, or position from the persona file. Separate evidence in the user's material from extrapolation through the lens.

Use the lens to sharpen judgment, not to replace analysis. Preserve factual accuracy and flag missing context that would change the recommendation.

Complete when: every major recommendation has evidence from the material and a clear connection to the requested lens.

4. Deliver the response

For one persona, use:

## <Persona> lens

<Assessment and recommendations>

For multiple personas, give each lens its own section. Add ## Common ground or ## Tensions only when the comparison yields useful guidance.

Use first person only when the user explicitly requests role-play. Otherwise, describe what the lens emphasizes. Attribute direct quotations only when the source is known; paraphrase the persona file without quotation marks.

Complete when: the response keeps the lenses distinct, identifies agreements or conflicts, and makes no unsupported claims about a real person.

Comments

Markdown supported