---
title: "cave-talk"
description: "Ultra-compressed communication mode. Cuts output tokens while keeping full technical accuracy."
canonical: https://agentpluginsdirectory.com/plugins/cave-talk
last-updated: 2026-10-05
---

# cave-talk
Ultra-compressed communication mode. Cuts output tokens while keeping full technical accuracy.
- Slug: cave-talk
- Publisher: kilianpaquier
- Repository: https://github.com/kilianpaquier/ai-integration
- Manifest: plugins/cave-talk/plugin.json
- Version: 0.7.0
- License: MIT and Apache-2.0
- Category (editorial): other
- Skills: 0
- MCP servers: 0
- Stars: 0
- Repository created: 2026-07-19
- Repository last pushed: 2026-10-04
- Publisher type: User
- Listing: https://agentpluginsdirectory.com/plugins/cave-talk
- Schema: https://agent-plugins.org/schemas/1.0.0/plugin.schema.json

## What cave-talk does, in the publisher's words

Ultra-compressed communication mode. Cuts output tokens while keeping full technical accuracy.

Uses Caveman, with no affiliation or endorsement from its author.

Caveman supports the following levels: caveman, ultracave, megacave, or off to disable. Within a session, /caveman switches the level and stop caveman (or normal mode) turns it off.

- the CAVEMAN_DEFAULT_MODE environment variable
- a.caveman/config.json or.caveman.json within a repository with defaultMode property
- a ~/.config/caveman/config.json (or under $XDG_CONFIG_HOME/caveman/config.json) with defaultMode property

From the project README, punctuation lightly normalized. Full text: https://raw.githubusercontent.com/kilianpaquier/ai-integration/HEAD/plugins/cave-talk/README.md