This skill should be used when the user invokes "/discover", asks to "find patterns", "extract workflows", "what could be a skill", "what could be a command", "modularize this", "discover repeatable procedures", or wants to identify engineering patterns that could become new plugin components. Provides methodology for subsystem-aware pattern extraction and component type selection.
Analyzes engineering sessions to identify repeatable patterns and determine optimal component types for modularization.
npx claudepluginhub mberto10/mberto-compoundThis skill inherits all available tools. When active, it can use any tool Claude has access to.
discovery-craftAnalyze engineering work sessions to discover modularizable patterns and determine the right component type: skill, command, agent, or hook. This is subsystem-aware pattern extraction — grounding discovery in concrete subsystem knowledge rather than abstract observation.
Load references/compounding-methodology.md for the underlying philosophy — the compound loop, what makes learning worth encoding, and the frontier model.
Every repeatable procedure with clear inputs and outputs can become a loadable module. The question:
"What procedure lives in people's heads that could be a plugin component?"
In compound engineering, extend this to:
"What subsystem-specific knowledge keeps being needed but isn't captured?"
Use when: Pattern provides specialized knowledge or methodology that Claude should apply when relevant.
Characteristics:
In engineering context: Domain knowledge about a subsystem, design patterns used in the codebase, testing strategies for specific areas.
Use when: Pattern is a user-initiated action with clear steps.
Characteristics:
/command-nameIn engineering context: Deployment procedures, migration workflows, code generation for specific patterns.
Use when: Pattern requires autonomous, multi-step work that benefits from specialization.
Characteristics:
In engineering context: Code review from a specific angle, dependency analysis, test generation.
Use when: Pattern should happen automatically in response to events.
Characteristics:
In engineering context: Pre-commit checks, invariant validation, file change guards.
| Pattern Characteristic | Component |
|---|---|
| Knowledge Claude should know | Skill |
| Action user explicitly requests | Command |
| Autonomous specialized subtask | Agent |
| Should happen on every X event | Hook |
| Combines knowledge + action | Command + Skill |
| Complex workflow with subspecialties | Command + Agents |
Examine what happened with subsystem context:
helpful_skills have gaps — knowledge that was needed but not available?For each pattern, document:
Pattern: [Name]
Trigger: When would this be needed?
Inputs: What does it start with?
Process: What happens (3-7 steps)?
Outputs: What does it produce?
Knowledge: What expertise is embedded?
Subsystems: Which subsystems does this relate to?
Frequency: How often does this come up?
Apply the decision matrix:
[ ] Is this knowledge Claude should apply contextually?
→ Yes: SKILL (add to subsystem helpful_skills)
[ ] Is this an action users explicitly request?
→ Yes: COMMAND (add to local project plugin)
[ ] Is this autonomous work for a specialist?
→ Yes: AGENT (add to local project plugin)
[ ] Should this happen automatically on events?
→ Yes: HOOK (add to local project plugin)
[ ] Multiple apply?
→ Design combination
| Criterion | Question |
|---|---|
| Recurrence | Will this happen again in this project? |
| Clear boundaries | Are inputs/outputs defined? |
| Teachable | Can it be written as instructions? |
| Consistent | Should it work the same way each time? |
| Value | Is packaging worth the effort? |
For each proposed component:
helpful_skills sections to update.Features built for one purpose often work for another. For each discovered pattern:
Don't over-specialize. Look for the general capability.
## Proposed Skill: [name]
**Location:** .claude/plugins/{project-plugin}/skills/[skill-name]/
### Trigger Description
This skill should be used when [triggers].
### What It Provides
- [Knowledge area]
- [Workflow guidance]
### Subsystem Integration
- Add to helpful_skills in: [subsystem1, subsystem2]
- Addresses gap: [GAP-XXX if applicable]
### Estimated Size
- SKILL.md: ~[X] words
## Proposed Command: /[name]
**Location:** .claude/plugins/{project-plugin}/commands/[name].md
### Frontmatter
description: [one-line description]
argument-hint: [arguments]
allowed-tools: [tools needed]
### Workflow
1. [Step]
2. [Step]
3. [Step]
### Subsystem Integration
- Reads from: [subsystem specs]
## Proposed Agent: [name]
**Location:** .claude/plugins/{project-plugin}/agents/[name].md
### When to Use
[Situation where this agent helps]
### Expertise
[What it's good at]
### Tools Required
- [Tool]: [why]
## Proposed Hook: [name]
**Location:** .claude/plugins/{project-plugin}/hooks/hooks.json
### Event Type
[PreToolUse | PostToolUse | Stop | etc.]
### Purpose
[What it enforces/validates]
### Configuration
{JSON config}
Wrong component type:
Over-engineering:
Under-engineering:
Discovery produces component specifications ready for implementation via /consolidate. Each spec should be complete enough to build from.
Creating algorithmic art using p5.js with seeded randomness and interactive parameter exploration. Use this when users request creating art using code, generative art, algorithmic art, flow fields, or particle systems. Create original algorithmic art rather than copying existing artists' work to avoid copyright violations.
Applies Anthropic's official brand colors and typography to any sort of artifact that may benefit from having Anthropic's look-and-feel. Use it when brand colors or style guidelines, visual formatting, or company design standards apply.
Create beautiful visual art in .png and .pdf documents using design philosophy. You should use this skill when the user asks to create a poster, piece of art, design, or other static piece. Create original visual designs, never copying existing artists' work to avoid copyright violations.