ies/music-topos/.claude-marketplaces/topos-skills/plugins/topos-skills/skills/three-match/SKILL.md
3-MATCH gadgets and non-backtracking geodesics for 3-SAT via colored subgraph isomorphism.
npx skillsauth add plurigrid/asi three-matchInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Status: ✅ Production Ready Trit: -1 (MINUS - conservative/geodesic) Principle: Local constraints → Global correctness Frame: Non-backtracking geodesics with Möbius filtering
Three-Match reduces 3-SAT to 3-coloring which reduces to colored subgraph isomorphism. The 3-MATCH gadget enforces constraints LOCALLY via:
Correct by construction: If local geodesic constraints are satisfied, global 3-SAT solution is guaranteed.
# Three colors match at depth d iff:
# - Pairwise differences have 3-adic valuation ≥ d
# - No backtracking (each color unique in path)
# - GF(3) sum ≡ 0 (mod 3)
v₃(|a - b|) ≥ d ∧ v₃(|b - c|) ≥ d ∧ v₃(|c - a|) ≥ d
Three colors forming a valid local constraint:
match = ThreeMatchGeodesicGadget::ThreeMatch.new(seed: 0x42D, depth: 1)
match.color_a # => { trit: -1, hex: "#2626D8", polarity: :minus }
match.color_b # => { trit: 0, hex: "#26D826", polarity: :ergodic }
match.color_c # => { trit: 1, hex: "#D82626", polarity: :plus }
match.gf3_conserved? # => true
Prime path through color space:
geo = NonBacktrackingGeodesic.new(seed: seed, length: 8).generate!
geo.prime? # => true (no backtracking)
geo.moebius_product # => ±1 (non-zero for primes)
geo.moebius_filter # => filtered path (only primes kept)
3-SAT clause reduction:
gadget = ColoredSubgraphGadget.new(seed: seed)
gadget.add_clause(1, -2, 3) # (x₁ ∨ ¬x₂ ∨ x₃)
gadget.add_clause(-1, 2, 4) # (¬x₁ ∨ x₂ ∨ x₄)
gadget.build_gadgets!
gadget.correct_by_construction? # => true
Möbius inversion bidirectionally:
filter = BackAndForthFilter.new(seed: seed)
result = filter.full_cycle(sequence)
# Primes kept, composites filtered
# Run 3-MATCH demo
just three-match
# Test gadget correctness
just test-three-match
# Combine with unworld
just unworld-match
require 'three_match_geodesic_gadget'
# Create gadget
match = ThreeMatchGeodesicGadget::ThreeMatch.new(seed: seed)
# Verify constraints
match.gf3_conserved? # GF(3) sum = 0
match.matches_at_depth?(1) # 3-adic valuation ≥ 1
# Build geodesic
geo = ThreeMatchGeodesicGadget::NonBacktrackingGeodesic.new(
seed: seed, length: 12
).generate!
# Check primality
geo.prime? # No backtracking?
The 3-MATCH chain uses seed-chaining for gadget sequence:
chain = Unworld::ThreeMatchChain.new(genesis_seed: seed, length: 4)
chain.unworld[:matches].each do |m|
puts "#{m[:colors]} | GF(3): #{m[:gf3]}"
end
μ(n) = { 1 if n = 1
{ (-1)^k if n = p₁p₂...pₖ (distinct primes)
{ 0 if n has squared prime factor
f(n) = Σ_{d|n} g(d) ⟹ g(n) = Σ_{d|n} μ(n/d) f(d)
v₃(n) = max { k : 3^k | n }
─── 3-MATCH Gadget ───
3-MATCH(d=1): #D8267F #2CD826 #4FD826
GF(3) conserved: true
Matches at depth 1: true
─── Non-Backtracking Geodesic ───
Geodesic(PRIME, μ=1): #D8267F → #2CD826 → #4FD826 → ...
Prime path: true
Möbius product: 1
─── Colored Subgraph Gadget (3-SAT) ───
Clauses: 3
GF(3) all conserved: true
Prime geodesics: 3
Correct by construction: true
Skill Name: three-match Type: 3-SAT Reduction / Colored Subgraph Isomorphism Trit: -1 (MINUS) GF(3): Conserved by construction Geodesics: Non-backtracking (prime paths only)
development
BDD-Driven Mathematical Content Verification Skill Combines Behavior-Driven Development with mathematical formula extraction, verification, and transformation using: - Cucumber/Gherkin for specification - RSpec for implementation verification - mathpix-gem for LaTeX/mathematical content extraction - Pattern matching on syntax trees for formula validation Enables iterative discovery and verification of mathematical properties through executable specifications.
tools
Meta-skill that generates domain-specific AI skills from tool documentation
development
Code Query with AI-enhanced deterministic analysis via SplitMix ternary classification
development
Directed Yoneda lemma as directed path induction. Riehl-Shulman's key insight for synthetic ∞-categories.