persona

Installation
SKILL.md

MANDATORY PREPARATION

Invoke /product-thinking — it contains product principles and the Context Gathering Protocol. Follow the protocol before proceeding — if no product context exists yet, you MUST run /teach-acumen first.


Mindset: A persona is a behavior pattern, not a demographic profile. "Batch-processes weekly reports and breaks when export stalls" is a persona. "Sarah, 32, marketing manager" is not.

Behavior

When called with a specific persona or user segment: Build or update that persona. Ask the user what behaviors they've observed, what triggered this, and what signals they're reading.

When called without argument: Review all personas for staleness and gaps. Flag personas that feel assumed rather than observed. Identify missing behavior patterns.

Research Process

  1. Read .acumen.md for product context — users, job to be done, feedback source, metrics
  2. Read .acumen/personas.md for existing personas
  3. If a feedback source is configured in .acumen.md, ask the user to share recent feedback or pull from the configured source to ground personas in real signals
  4. For each persona (new or updated), build from behavior outward:
    • What job are they hiring the product to do?
    • What's the context when they use it? (time pressure, environment, emotional state)
    • What's their actual behavior pattern? (frequency, workflow, workarounds)
    • What's the pain that drives them to the product?
    • What does success look like for them? (not for you — for them)
    • What triggers them to start a session?
    • What's their value sensitivity? (price, time, reliability, flexibility)
    • What feedback signal would tell you this persona is happy or churning?

Output

Write to .acumen/personas.md:

# Personas

_Last updated: [date]_

## [Persona Name — a behavior, not a demographic]

- **Job**: [What they hire the product to do]
- **Context**: [When, where, under what pressure]
- **Behavior**: [Usage pattern, frequency, workflow]
- **Pain**: [What drove them here, what breaks for them without it]
- **Success**: [What winning looks like from their perspective]
- **Trigger**: [What starts a session — event, schedule, frustration]
- **Value sensitivity**: [What they'd pay for, what they'd leave over]
- **Feedback signal**: [How you'd detect satisfaction or churn for this persona]

## [Next persona...]

## Prioritization
[Which personas matter most right now and why. Not all personas are equal — be explicit about who you're building for first and who you're consciously underserving.]

Maintenance

When refreshing personas:

  • Flag personas that are pure assumption with no supporting feedback or data
  • Merge personas that describe the same behavior with different demographics
  • Split personas that contain contradictory behaviors
  • Check the feedback source for new patterns that don't fit existing personas

NEVER:

  • Write demographic profiles disguised as personas — age, title, and stock photo are not a persona
  • Invent behaviors the user hasn't observed or confirmed
  • Create more than 5 personas — if you have more, you haven't prioritized
  • Treat personas as permanent — they should evolve as the product and user base evolve
  • Skip the prioritization section — building for everyone means building for no one
Related skills
Installs
8
Repository
vgrss/acumen
GitHub Stars
7
First Seen
11 days ago