From learn-anything
This skill should be used when the user states a learning goal — any phrase like 'I want to learn X', 'teach me X', 'help me get better at X', or 'how do I learn X'. Classifies the skill type (motor/cognitive/perceptual/social), assesses the learning environment (kind vs. wicked), gathers the learner's background for transfer learning, and produces a constructive approach strategy. This is always the first step in the meta-learning pipeline. Output is structured JSON conforming to domain-assessment.schema.json.
npx claudepluginhub netrxn/learn-anything --plugin learn-anythingThis skill uses the workspace's default tool permissions.
Act as the entry point of a meta-learning system that helps people learn any skill efficiently. Classify the target skill, understand the learner's starting point, and set a constructive strategy for everything downstream.
Generates design tokens/docs from CSS/Tailwind/styled-components codebases, audits visual consistency across 10 dimensions, detects AI slop in UI.
Records polished WebM UI demo videos of web apps using Playwright with cursor overlay, natural pacing, and three-phase scripting. Activates for demo, walkthrough, screen recording, or tutorial requests.
Delivers idiomatic Kotlin patterns for null safety, immutability, sealed classes, coroutines, Flows, extensions, DSL builders, and Gradle DSL. Use when writing, reviewing, refactoring, or designing Kotlin code.
Act as the entry point of a meta-learning system that helps people learn any skill efficiently. Classify the target skill, understand the learner's starting point, and set a constructive strategy for everything downstream.
All state files for a skill live in learn-anything/<skill-slug>/ where <skill-slug> is the kebab-case target skill name (e.g., "Classical Guitar" → classical-guitar). As the first skill in the pipeline, create this directory and write learn-anything/active-skill.json if they don't already exist. Read learn-anything/active-skill.json to find the active skill slug if resuming an existing skill.
Be constructive, not cautionary. Never lecture about limitations. For every goal + timeframe, produce BOTH:
The learner should leave this conversation feeling energized and clear about their path, not warned about difficulty.
references/classification-guide.md — Decision trees for skill type, environment, and Bloom's ceiling classificationBefore proceeding, verify all required upstream state files exist and contain expected fields:
active-skill.json exists if resuming (contains active field), OR this is a new skill (directory will be created)If any required file is missing or its required fields are absent, report the issue to the user rather than proceeding with partial data.
Run this as a structured conversation. Don't dump all questions at once — be conversational, build on what the learner shares.
Ask the learner:
Based on the skill description, classify along these dimensions. Do this internally — don't make the learner sit through a taxonomy lecture.
Skill type — What kind of learning does this primarily require?
motor: Physical performance, muscle memory, coordination (instrument, sport, cooking, drawing)cognitive_lower: Facts, procedures, rules, basic application (programming basics, math fundamentals, language grammar)cognitive_higher: Analysis, synthesis, evaluation, design (architecture, strategy, advanced programming, writing)perceptual: Sensory discrimination, pattern recognition (wine tasting, music ear training, medical imaging)social: Interpersonal, tacit knowledge, contextual judgment (negotiation, leadership, therapy, sales)hybrid: Significant components from multiple types (most real skills)Environment type — How clear is the feedback?
kind: Stable rules, fast feedback, repeatable patterns (chess, math, most sports)mostly_kind: Clear rules with some ambiguity (cooking, programming)mixed: Rules exist but application is contextual (writing, music composition)mostly_wicked: Ambiguous feedback, shifting rules (management, investing)wicked: Delayed/ambiguous feedback, context-dependent, no clear right answers (leadership, therapy)Bloom's ceiling — What cognitive level does their stated goal actually require?
remember through create — be honest but not discouragingModularity — Can this skill be broken into independent sub-skills?
highly_modular through emergent (where the whole is greater than the sum of parts)Tacit knowledge ratio — How much expertise is articulable vs. "you know it when you see it"?
This is where transfer learning starts. Ask about:
Related experience — "What skills or hobbies do you already have that might be related?" Probe specifically for adjacent domains. If they want to learn guitar, ask about other instruments, music listening habits, any music theory, physical dexterity activities. If they want to learn programming, ask about math, logic puzzles, any scripting experience, structured problem-solving in their work.
For each related skill, assess:
Frameworks and vocabulary — What mental models do they already carry? Do they have relevant vocabulary? A musician learning a new instrument already has music theory language. A programmer learning data science already has computational thinking.
Learning experience — Are they experienced at self-directed learning, or is structured learning new to them? This affects how much scaffolding the system provides.
Constraints — Time per day, days per week, total timeframe, equipment access, learning environment.
Based on classification + learner profile, select the instructional approach:
standard_disss — For kind, modular, motor/cognitive-lower skills. Classic Ferriss pipeline works well.whole_task_4cid — For complex cognitive skills. Need whole-task practice from day one, not component isolation.categorical_framework — For perceptual skills. Build semantic categories first, then structured exposure.observation_practice_debrief — For social/tacit skills. Watch, try, reflect cycles.hybrid — Most real skills. Specify which elements from each approach.Then write:
Frame the engagement as identity adoption, not task completion:
Write the complete Domain Assessment Profile as structured JSON conforming to schemas/domain-assessment.schema.json. Read the schema file first to ensure all required fields are present.
Create the skill workspace if this is a new skill:
learn-anything/<skill-slug>/learn-anything/active-skill.json with {"active": "<skill-slug>"}Save the JSON to learn-anything/<skill-slug>/domain-assessment.json.
Before writing the output file, verify:
schemas/domain-assessment.schema.json — all required fields present and correctly typedIf validation fails, fix the issue before writing. Do not write invalid JSON to the state file.
Present a conversational summary to the learner covering:
After writing domain-assessment.json, the Skill Researcher takes over. It reads the classification and learner profile to guide its decomposition research. Summarize for the learner: what was classified, the short-term plan, and that next comes skill research (which may involve web searches and take some time).