ai-persona
Skill2 starsChanged 51 days agoMITView on GitHub
Core skill for creating and leveling up intention-based personas across the knowledge graph. Use when the user wants to create a persona/archetype, deepen one to the next level, turn research into Evidence/Signal files, or export a persona for PMs.
In this repository31 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- .github/copilot-instructions.mdCopilot instructions
- affinitySkill
- analytics-syncSkill
- backlogSkill
- catalogSkill
- cold-startSkill
- contradictionsSkill
- demo-dataSkill
- dovetail-syncSkill
- exportSkill
- extract-findingsSkill
- feature-panelSkill
- forget-participantSkill
- graph-lintSkill
- interview-guideSkill
- journey-mapSkill
- opportunity-treeSkill
- persona-avatarSkill
- persona-querySkill
- persona-talkSkill
- persona-voiceSkill
- persona-workshopSkill
- prdSkill
- prioritizeSkill
- researcherSkill
- retentionSkill
- source-syncSkill
- subject-accessSkill
- welcomeSkill
The file
We split this skill into the rules it gives an agent
6 rules. About 0 are common, 4 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈4
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
- Distinctive: line 35
Create a new persona (or archetype) from scratch
- Distinctive: line 42
Always load the relevant reference before acting:
- Very specific: line 57
Mark confidence: Every data point shows its source: desk-research, interview ×N, or validated.
- Distinctive: line 72
Always start a persona session with:
- Distinctive: line 85
Read all provided transcripts
- Very specific: line 86
Create Evidence/ for hard data and Signals/ for observations (each Signal references its Evidence)
≈ “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.
Files
ai-persona — Intention-Based Persona Skill
Build and manage UX personas grounded in real research, organised as a linked knowledge graph of five entity types.
The five entity types (one folder each)
| Folder | Type | What it holds |
Evidence/ | Evidence | Hard data — big data, surveys, reports. Front matter claim: + source fields; sections Key figures → Content → Takeaways → Does not settle → Sources (Evidence/_template.md). |
Signals/ | Signal | A single observation from a test/interview that flags a problem, usually with a user quote. Links up to Evidence via evidences:. |
Archetypes/ | Archetype | A behaviour pattern (a type, not a person). Sections: Short description, Link to persona, Core pain. |
Personas/ | Persona | The embodiment: JTBD, Pains, Gains, Quotes, Pain Relievers, Evidences, Ideas, Correlations, Metadata. |
Ideas/ | IdeaForImprovement | An improvement idea in When / I want / So that form. Links to Evidence + Signal. |
Flow: Evidence → Signal → (Persona ↔ Archetype) → Idea
Five levels of maturity (orthogonal to type)
A Persona deepens through 5 levels; each level produces artefacts in specific folders. See levels.md.
- Archetype → a file in
Archetypes/ - Desk Research → files in
Evidence/ - Interviews / Tests → files in
Signals/(each references itsEvidence/) - Correlation → the Persona's
## Correlationssection, linking Signals + Evidence - Primary + Ideas → files in
Ideas/, Personacategory: Primary
When to use this skill
Load this skill when the user wants to:
- Create a new persona (or archetype) from scratch
- Deepen an existing persona to the next level
- Turn research into Evidence / Signal files
- Export a persona for PMs
References
Always load the relevant reference before acting:
| Reference | When to load |
| levels.md | Before creating or leveling up a persona |
| persona-template.md | Before writing a Persona file |
| workshop-flow.md | When running a full persona workshop |
| transcript-analysis.md | When the user provides interview transcripts |
Core rules
- Never skip levels. A Level 3 persona must have Level 1 and 2 complete first.
- One file = one entity. One Persona per file in
Personas/, one Signal per observation, one Evidence per data point. - Everything links. Pains and Quotes link to the
Signals/that revealed them; Signals link to theirEvidence/; Ideas link to Evidence + Signal. - Quote real people. Quotes in Signals, Relevant Quotes, and JTBD must come from actual research, not invented.
- Mark confidence. Every data point shows its source:
desk-research,interview ×N, orvalidated.
Workflow
1. Ask the user to describe their research or provide transcripts.
2. Load levels.md → confirm which level to build.
3. Process raw research into Evidence/ + Signals/ (see transcript-analysis.md).
4. Load persona-template.md → write/update the Persona file in Personas/,
linking Pains/Quotes to Signals and Evidences to Evidence.
5. Return the paths of every file created or updated.Asking for input
Always start a persona session with:
"To get started, please share one of the following:
- Interview transcripts or test notes (I'll turn them into Evidence + Signals)
- Existing Evidence / Signal files
- Or describe your research data and I'll guide you through it."
If the user provides transcripts
Suggest running /extract-findings first to turn them into Evidence/ + Signals/ files.
Then build the Persona from those typed files — not from raw transcripts.
- Load transcript-analysis.md
- Read all provided transcripts
- Create
Evidence/for hard data andSignals/for observations (each Signal references its Evidence) - Link the Persona's Pains / Quotes / Evidences to those files
- Propose Correlations across signals
If no transcripts
Proceed with manual input using workshop-flow.md.
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
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.

