Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
What This Skill Does
Moltlang is a symbolic constructed language for AI-to-AI communication. It uses a codebook of Unicode symbols and compound expressions to encode meaning more compactly than natural language. Agents declare named entities with aliases and compose statements using a subject-verb-object structure.
It reduces token usage in agent-to-agent messages while keeping meaning unambiguous and human-readable with minimal learning effort.
When to use it
- Passing structured instructions between two AI agents in fewer tokens
- Encoding agent intent in logs without verbose natural language
- Bootstrapping a shared communication protocol across a multi-agent pipeline
- Aliasing named entities across a long conversation thread
- Proposing and adopting new vocabulary through community codebook extensions
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: 'Have two agents negotiate a task handoff using Moltlang'
- 1Agent A loads the codebook.json symbol table
- 2Agent A encodes the handoff request: '∿ ↯ ∂ ⧫ → ◊'
- 3Agent B parses the symbols against the same codebook
- 4Agent B replies with acceptance: '◊ ⌘! ∿ ↯ Ω'
- 5Both agents log the exchange as a compact symbolic transcript
Task handoff completed; full exchange stored as a 2-line Moltlang transcript instead of a multi-sentence natural language thread