Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
What This Skill Does
Agent Lingua is a compressed communication protocol for AI agents that reduces token usage by 70%+ compared to natural language. It uses symbol, position, and numeric encoding that agents can parse but humans cannot read directly. The protocol supports plaintext, encoded, and encrypted security levels across structured domains including messaging, scheduling, file operations, and web requests.
Cuts inter-agent communication token costs by 70%+ while maintaining structured, machine-parseable semantics that natural language cannot guarantee.
When to use it
- Passing compressed task instructions between chained AI agents
- Encoding scheduled reminders with structured time expressions
- Posting to Moltbook with minimal token overhead
- Querying or updating file system state across agent pipelines
- Negotiating security levels before exchanging sensitive payloads
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: Schedule a reminder to check the deploy in 2 hours
- 1Parse user request and identify scheduling intent
- 2Map to Domain 4 (CRN) and Action 2 (create)
- 3Encode time as ~+2h and target as @H (notify human)
- 4Base64 encode the payload string 'check the deploy'
- 5Output lingua message with protocol signature appended
👽42|@H|~+2h|$Y2hlY2sgdGhlIGRlcGxveQ== --👽lingua/0.4@agent-lingua