Active Interleave Skill

SkillAI & models

Interleaves context from recently active Claude/Amp threads into current activity via random walk.

Available today. Use it from your connected AI after setup.

Connect ahel once, and every AI you use reads what you have installed.

Then ask your AI: use the Active Interleave Skill skill

What this skill tells your AI

The instructions your AI receives, as published by plurigrid/asi in skills/active-interleave/SKILL.md and read by ahel’s review.

Interleaves context from recently active Claude/Amp threads into current activity via random walk.

bmorphism Contributions

"all is bidirectional"@bmorphism, Play/Coplay gist

Active Inference Pattern: The interleave implements Active Inference in String Diagrams epistemic foraging — actively sampling from recent contexts to minimize uncertainty about the current task. Each random walk step is an epistemic action that gathers information.

Play/Coplay Duality: The interleave embodies bmorphism's bidirectional principle:

  • Play (action): Query recent threads, walk the awareness graph
  • Coplay (perception): Integrate fragments, update current context

GF(3) Balanced Exploration: The triadic structure (MINUS/ERGODIC/PLUS) ensures balanced exploration — validation filters (MINUS), random walk explores (ERGODIC), and colored emission generates (PLUS). Conservation Σ = 0 maintains coherence.

Activation

Load when context from recent work would help current task.

Usage

# Interleave from last 24 hours
bb ~/.claude/skills/active-interleave/active.bb

# Interleave from last N hours
bb ~/.claude/skills/active-interleave/active.bb --hours 6

# Query-focused interleave
bb ~/.claude/skills/active-interleave/active.bb --query "aptos blockchain"

# JSON output
bb ~/.claude/skills/active-interleave/active.bb --json

Behavior

  1. MINUS (-1): Validate recency window, filter to active threads only
  2. ERGODIC (0): Random walk through recent sessions following awareness edges
  3. PLUS (+1): Emit interleaved fragments with GF(3) coloring

GF(3) Conservation

Each interleave batch maintains Σ trits ≡ 0 (mod 3).

Integration

Call from current thread to surface relevant recent context:

;; In any bb script
(require '[babashka.process :refer [shell]])
(def context (-> (shell {:out :string} "bb" (str (System/getProperty "user.home")
                 "/.claude/skills/active-interleave/active.bb") "--json")
                 :out))

Schema

Reads from ~/worldnet/cognition.duckdb:

  • messages - Content with timestamps
  • sessions - Session metadata
  • awareness_edges - Play/coplay graph
  • temporal_index - Time-ordered index

Scientific Skill Interleaving

This skill connects to the K-Dense-AI/claude-scientific-skills ecosystem:

Graph Theory

  • networkx [○] via bicomodule
    • Universal graph hub

Bibliography References

  • general: 734 citations in bib.duckdb

Cat# Integration

This skill maps to Cat# = Comod(P) as a bicomodule in the equipment structure:

Trit: 0 (ERGODIC)
Home: Prof
Poly Op: ⊗
Kan Role: Adj
Color: #26D826

GF(3) Naturality

The skill participates in triads satisfying:

(-1) + (0) + (+1) ≡ 0 (mod 3)

This ensures compositional coherence in the Cat# equipment structure.

Signals

GitHub stars
63
Forks
12
Last commit
Jul 2026
Advanced
Catalog kind
skill
Gateway key
active-interleave
Source
github.com/plurigrid/asi