growth-psych
Skill4 starsChanged 10 days agoMITView on GitHub
Audit landing, pricing and signup pages, review multi-step flows, and design growth experiments using 106 named psychology principles. Use for CRO or UX audits, flow reviews, A/B test ideas, "why isn't this converting", or a URL to critique.
In this repository15 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- cavemanSkill
- gh-issues-by-severitySkill
- grid-helpSkill
- grill-meSkill
- handoffSkill
- mine-learningsSkill
- openspec-helpSkill
- rubber-duckSkill
- setup-repo-skillsSkill
- skill-scoutSkill
- spec-scoutSkill
- standupSkill
- tightenSkill
- un-claudishSkill
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 1 is distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
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
- Very specific: line 22
Context: If .grid/project.yaml exists, read it for brand, audience, funnel and metrics.
- Very specific: line 27
Principles: Read only the categories the step needs: sections/principles-information.md (perception, layout, actions), -meaning.md (proof, …
- Distinctive: line 39
Read a section only when its condition applies.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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
growth-psych
Conversion work anchored to named psychology principles. A finding without a principle is an opinion; one without a fix is a complaint. Serves lead or enquiry forms, self-serve signup and direct purchase.
Principle catalogue adapted from growth.design/psychology (free cheat-sheet PDF linked there). Names and four categories follow it; every definition here is our own wording.
Modes
Pick from the request; ask if unclear.
- audit: teardown of an existing page (landing, pricing, signup, campaign). Read
sections/audit-zones.md, thensections/deliverable.md. - flow: step-by-step review of a multi-step flow or mock (signup, onboarding, checkout, enquiry, booking), before or after build. Read
sections/flow-review.md, thensections/deliverable.md. - experiment: turn a principle into a testable hypothesis with metric, sample size and stop rule. Read
sections/experiments.md.
Core (all modes)
- Context. If
.grid/project.yamlexists, read it for brand, audience, funnel and metrics. Never assume them; ask for the goal and audience if absent. - Visitor and job. Who arrives, from where, wanting what; what the page or flow wants (lead sent, account created, payment made). Competing goals are a finding.
- Finding = observation + principle + why + fix. Quote the real copy or describe the real element. Name the principle with its one-line meaning. Say why it matters for this goal. Write the actual replacement copy, not "be more benefit-driven".
- Balance. Report principles used well, violated and missing. Depth over breadth: 8-15 findings.
- Rate each finding: Impact 🔴 high / 🟠 medium / 🟡 low; Effort 🟢 copy or setting / 🟡 new asset / 🔴 redesign. Quick wins first.
- Principles. Read only the categories the step needs:
sections/principles-information.md(perception, layout, actions),-meaning.md(proof, offer, trust),-time.md(pricing, friction, commitment),-memory.md(endings, delight, retention).
Capture
Fetch copy and structure with WebFetch. For visual claims use a browser: browse (gstack) or anthropic-skills:chrome-browser, at desktop and mobile widths. With neither, work from fetched copy and say "visual claims unverified". User screenshots win for visuals.
Ethics
Persuade, never trick. Flag as trust findings, even if they convert: fake or resetting scarcity, pre-ticked opt-ins, confirmshaming, hidden costs, obstructed exits. When a manipulative pattern would work, say so and give the ethical alternative.
Sections
Read a section only when its condition applies; do not preload them.
- sections/audit-zones.md: read when running audit mode (zone checklist, two passes)
- sections/flow-review.md: read when running flow mode
- sections/experiments.md: read when running experiment mode or turning findings into tests
- sections/deliverable.md: read when building an audit or flow report (structure, scoring, HTML)
- sections/principles-information.md: read when checking perception, hierarchy, layout, calls to action
- sections/principles-meaning.md: read when checking proof, offer, positioning, trust, copy
- sections/principles-time.md: read when checking pricing, defaults, friction, commitment, urgency
- sections/principles-memory.md: read when checking endings, onboarding, delight, retention
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.

