aeromomoaeromomoCommunityยท1 Agent Skills Included

claw-compactor

Compress LLM prompts and agent memory to cut token costs

Compresses code, JSON, logs, diffs, and chat history through a 14-stage pipeline to shrink LLM token usage by 15-82%. Eliminates wasted context window space and rising API bills without losing meaning, using reversible compression and AST-aware code handling. Includes Engram, a real-time memory layer that turns long agent conversations into structured summaries for persistent context.
npx skills add aeromomo/claw-compactor --all -g -y

All Skills in This Repository (1)

Pure Emerald Level Indicators

Frequently Asked Questions

FAQPage Schema
How to install Claw Compactor?โ–ผ

Run `npx skills add aeromomo/claw-compactor --all -g -y` in your terminal to install the full skill suite globally.

How to reduce LLM token costs?โ–ผ

Claw Compactor runs your prompts and workspace files through 14 compression stages, cutting token usage by 15-82% while keeping the content readable for the model.

Does compression lose important information?โ–ผ

No. It uses reversible compression with a rewind store, so the AI can retrieve any original section on demand, and code identifiers are never altered.

What is Engram observational memory?โ–ผ

Engram is a live memory layer that watches agent conversations and compresses them into priority-tagged summaries, giving long-running agents persistent long-term context.

Does Claw Compactor work with OpenClaw agents?โ–ผ

Yes. It ships as a standard SKILL.md skill that OpenClaw agents trigger automatically on phrases like compress memory or save tokens.

Related Repositories in Software Engineering

View All in Software Engineeringโ†’