install
source · Clone the upstream repo
git clone https://github.com/plurigrid/asi
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/plurigrid/asi "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skills/teglon-bison" ~/.claude/skills/plurigrid-asi-teglon-bison && rm -rf "$T"
manifest:
skills/teglon-bison/SKILL.mdsource content
Teglon Bison Skill
Bisimulation bistability from reversible software and conservative Logik
Source: TeglonLabs/bison (fork of unisoncomputing/unison-llm-support)
Trit: +1 (PLUS)
Substrate: Semantic (LLM Societies)
Overview
Bison extends Unison's content-addressed code with bisimulation semantics:
- Two programs are bisimilar if they can simulate each other step-by-step
- Reversible computation: every step can be undone
- Conservative logic: information is never destroyed
α/β/γ Diff Structure
| Arrow | Bison Meaning | Operation |
|---|---|---|
| α (−1) | Forward simulation step | |
| β (0) | Code hash mutation | |
| γ (+1) | Bisimulation proof | |
Unison Integration
-- Content-addressed function squareRoot : Float -> Float squareRoot x = ... -- The hash IS the identity #abc123.squareRoot
Bisimulation Game
Two processes P and Q are bisimilar (P ~ Q) iff:
- If P →ᵃ P', then ∃Q'. Q →ᵃ Q' and P' ~ Q'
- If Q →ᵃ Q', then ∃P'. P →ᵃ P' and P' ~ Q'
Upstream Diff
From TeglonLabs fork:
- Added: Bisimulation checker for LLM-generated code
- Added: Conservative logic constraints
- Modified: Hash comparison with semantic equivalence
End-of-Skill Interface
GF(3) Triads
narya-proofs (-1) ⊗ bisimulation-game (0) ⊗ teglon-bison (+1) = 0 ✓ three-match (-1) ⊗ just-monad (0) ⊗ teglon-bison (+1) = 0 ✓
Autopoietic Marginalia
The interaction IS the skill improving itself.
Every use of this skill is an opportunity for worlding:
- MEMORY (-1): Record what was learned
- REMEMBERING (0): Connect patterns to other skills
- WORLDING (+1): Evolve the skill based on use
Add Interaction Exemplars here as the skill is used.