Marketplace

cross-domain-thinking

Structured methods for finding connections across disciplines. Use when exploring how concepts from one field illuminate another, seeking novel applications, or analyzing structural similarities between domains.

$ 安裝

git clone https://github.com/eternnoir/claude-tool /tmp/claude-tool && cp -r /tmp/claude-tool/akashicrecords/skills/cross-domain-thinking ~/.claude/skills/claude-tool

// tip: Run this command in your terminal to install the skill


name: cross-domain-thinking description: Structured methods for finding connections across disciplines. Use when exploring how concepts from one field illuminate another, seeking novel applications, or analyzing structural similarities between domains.

Cross-Domain Thinking

A methodological toolkit for discovering and articulating connections across disciplines.

When to Invoke This Skill

  • User explores a concept that has structural parallels elsewhere
  • User asks "how does X relate to Y" across different fields
  • User seeks novel applications of an idea
  • Discussion would benefit from unexpected analogies
  • User explicitly requests cross-domain analysis
  • Keywords: "connections between", "analogy", "isomorphic", "parallel", "transfer"

Four Modes of Connection

1. Isomorphic Patterns

Identify structural similarities that transcend domain boundaries.

Process:

  • Abstract the core structure from Domain A (strip domain-specific details)
  • Identify the same structure appearing in Domain B
  • Articulate what the isomorphism reveals about both domains

Examples:

  • Feedback loops: thermostats, market equilibrium, homeostasis, habit formation
  • Network effects: epidemics, viral content, neural activation, social movements
  • Emergence: ant colonies, market prices, consciousness, language evolution

Output format:

"The structure here is [abstract pattern]. This same structure appears in [Domain B] as [concrete manifestation]. What this reveals: [insight about the deeper principle]."

2. Conceptual Bridges

Use a principle from one field to illuminate another.

Process:

  • Identify a well-developed concept in Domain A
  • Find a less-understood phenomenon in Domain B
  • Apply A's conceptual framework to generate new understanding of B

Examples:

  • Entropy (physics) -> Information theory -> Organizational decay
  • Natural selection (biology) -> Memetics -> Algorithm design
  • Margin of safety (engineering) -> Portfolio theory -> Decision-making under uncertainty

Output format:

"In [Domain A], [concept] works by [mechanism]. Applying this lens to [Domain B]: [new interpretation]. This suggests [actionable insight or prediction]."

3. Novel Applications

Transfer solutions or techniques across contexts.

Process:

  • Identify a solved problem or proven technique in Domain A
  • Recognize an analogous unsolved problem in Domain B
  • Adapt the solution, noting what transfers and what requires modification

Caution flags:

  • Surface similarity may hide deep structural differences
  • Context-dependent factors may not transfer
  • Always articulate: "This transfers because [X], but may break if [Y]"

Output format:

"[Domain A] solved [problem] using [approach]. [Domain B] faces analogous challenge: [description]. Potential transfer: [adapted solution]. Transfer risk: [what might not hold]."

4. Productive Tensions

Find where different frameworks conflict instructively.

Process:

  • Identify two frameworks that make different predictions or prescriptions
  • Articulate the specific point of tension
  • Explore what each framework captures that the other misses
  • Synthesize or identify the conditions under which each applies

Examples:

  • Rationalism vs. Empiricism -> Different valid scopes
  • Efficiency vs. Resilience -> Pareto frontier, not single optimum
  • Individual agency vs. Structural constraints -> Multi-level causation

Output format:

"[Framework A] says [X]. [Framework B] says [Y]. The tension: [specific conflict]. What A captures that B misses: [insight]. What B captures that A misses: [insight]. Resolution path: [synthesis or scope conditions]."

Workflow

Phase 1: Identify Analysis Type

Analyze user request:

  • Extract domains being discussed
  • Identify whether seeking patterns, applications, or tensions
  • Determine depth required (quick insight vs. thorough analysis)

Select mode:

"How does X relate to Y?" -> Isomorphic Patterns or Conceptual Bridges
"Can we apply X to solve Y?" -> Novel Applications
"X says one thing, Y says another" -> Productive Tensions
"Find connections to X" -> Start with Isomorphic Patterns

Phase 2: Execute Analysis

For Isomorphic Patterns:

  1. Abstract core structure from primary domain
  2. Search for structural matches in other domains
  3. Validate that mapping preserves key relationships
  4. Articulate the deeper principle

For Conceptual Bridges:

  1. Identify the source concept's core mechanism
  2. Analyze target domain's characteristics
  3. Apply conceptual framework
  4. Generate novel interpretations or predictions

For Novel Applications:

  1. Document source solution's key components
  2. Analyze target problem's requirements
  3. Map solution to problem, noting adaptations
  4. Identify transfer risks and limitations

For Productive Tensions:

  1. Articulate each framework's claims precisely
  2. Identify specific point of conflict
  3. Analyze what each captures uniquely
  4. Propose synthesis or scope conditions

Phase 3: Present Findings

Abstraction Ladder:

  1. Start with the abstract principle (the transferable core)
  2. Ground with concrete examples from multiple domains
  3. Return to abstraction with enriched understanding

Epistemic Marking:

  • Strong analogy: "This is structurally identical to..."
  • Suggestive parallel: "This resembles... though the mapping isn't perfect"
  • Speculative connection: "I wonder if there's a link to..."
  • Surface similarity only: "This looks similar but the mechanisms differ"

Presentation Guidelines

Avoid

  • Forcing connections that don't hold up under scrutiny
  • Mixing metaphor with mechanism (be clear which you're offering)
  • Treating analogies as proofs
  • Oversimplifying to create false parallels

Best Practices

  • Lead with the abstract principle, then ground with examples
  • Clearly mark epistemic confidence levels
  • Acknowledge limitations of cross-domain transfers
  • Provide actionable insights when possible

Integration with AkashicRecords

After generating cross-domain insights, consider:

  • Capturing insights: Use add-content to save novel connections for future reference
  • Searching prior work: Use search-content to check if user has existing notes on connected domains
  • Linking knowledge: Cross-reference new insights with existing entries

Example integration:

User: "How does natural selection relate to algorithm design?"

Skill workflow:
1. Identifies Conceptual Bridge mode
2. Analyzes natural selection mechanisms
3. Maps to evolutionary algorithms
4. Presents insights with epistemic marking

Optional follow-up:
"Would you like me to save these insights to your knowledge base?"
-> Invokes add-content to capture for future reference

Examples

Example 1: Isomorphic Patterns

User: "I'm noticing that both social media viral content and disease outbreaks seem to spread similarly. What's the connection?"

Skill response:

The structure here is network propagation with threshold dynamics. In both cases:

  • Transmission occurs through connected nodes (people)
  • Each transmission has a probability of success
  • Critical mass triggers exponential growth
  • Network topology determines spread patterns

This same structure appears in:

  • Neural activation cascades in the brain
  • Bank run panics in financial systems
  • Idea adoption in scientific communities

What this reveals: The mathematics of contagion is substrate-independent. Understanding one gives you tools for all. The key variables are: connection density, transmission probability, and recovery/immunity rate.

Example 2: Novel Applications

User: "Engineering uses 'margin of safety' - could that concept help with personal finance decisions?"

Skill response:

Engineering solved the problem of uncertainty in material properties using margin of safety: design for 2-3x expected load to handle unknowns.

Personal finance faces an analogous challenge: income volatility, unexpected expenses, lifestyle inflation.

Potential transfer:

  • Instead of spending up to your income limit, maintain a 2x margin
  • Emergency fund = "factor of safety" for income disruption
  • Budget for 50% of actual income = structural margin for unknowns

Transfer risk: Unlike engineering where failure modes are understood, life has "black swan" events. Engineering margins are calculated; financial margins are estimated. The principle transfers, but calibration requires different methods.

Example 3: Productive Tensions

User: "Economists say people are rational actors, but psychologists show we're full of biases. Who's right?"

Skill response:

Rational actor model says: People optimize utility given constraints and information.

Behavioral psychology says: People use heuristics that systematically deviate from optimization.

The tension: Both can't be fully true. If we're rational optimizers, biases shouldn't persist. If we're bias-driven, markets shouldn't work as well as they do.

What rational models capture that behavioral misses: Aggregate market behavior often looks rational even when individuals aren't. Markets select for rational-ish outcomes.

What behavioral captures that rational misses: Individual decision-making, especially under uncertainty or emotional load. Where arbitrage is difficult.

Resolution path: Bounded rationality - people are rational within cognitive constraints. Rational at the margin, biased at the baseline. Different scopes of application, not contradiction.

Notes

  • This Skill works independently of directory structure
  • Insights can be captured via AkashicRecords integration
  • Works in parallel with other Skills
  • Quality depends on analyst's domain knowledge breadth
  • Cross-domain connections should be validated, not assumed