skills.rest
snewhouse
snewhouseCOMMUNITY
Explore all skills by this author →
𝕏@s_j_newhouseGitHub

token-compression

Reduce AI agent output tokens with mode-based compression rules for AA-MA workflows.

6|1|Updated Apr 5, 2026
Search Tags:#Claude Code#cost reduction#token optimization#agent memory#autonomous execution#prompt compression#AA-MA
One-click install
npx skills add https://github.com/snewhouse/aa-ma-forge --skill token-compression-snewhouse
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: token-compression
Source: https://github.com/snewhouse/aa-ma-forge/tree/main/claude-code/skills/token-compression
Command: npx skills add https://github.com/snewhouse/aa-ma-forge --skill token-compression-snewhouse

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Token compression reduces the number of output tokens an AI agent produces while preserving full technical accuracy, helping multi-session workflows stay efficient and cost-effective.

Core Features & Use Cases

  • Mode-aware compression: Automatically maps AA-MA task Mode to compression intensity (HITL→lite, AFK→ultra, default→full) to match human review versus autonomous execution needs.
  • Manual override: Lets you explicitly choose lite, full, or ultra with /compress lite|full|ultra, or disable via /compress off.
  • Safety-oriented boundaries: Protects code blocks, commit messages, PR descriptions, and exact error text from compression changes, and avoids compression around destructive or irreversible actions.

Quick Start

Use token-compression to generate a tighter response by asking the agent to run the instruction: /compress full and then complete your request using the compressed style.

Frequently Asked Questions about token-compression

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I reduce AI agent output tokens without losing technical accuracy?▼

Token compression reduces AI agent output token count by applying structured compression rules for multi-session workflows, preserving full technical accuracy while cutting response length and cost.

How does token compression handle code blocks and error messages?▼

Token compression establishes safe boundaries that protect code blocks, commit messages, PR descriptions, and exact error text from compression changes to preserve technical accuracy.

Can I manually adjust token compression intensity for different workflows?▼

You can manually override compression intensity using commands like /compress lite, full, or ultra, or disable it entirely with /compress off to suit your workflow.

What is the best way to optimize token usage for autonomous execution?▼

Token compression maps autonomous AFK modes to ultra intensity, HITL modes to lite compression, and default modes to full, matching compression intensity to your execution needs.

Does token compression work with multi-session agent workflows?▼

Token compression is designed for multi-session execution workflows where response length impacts cost, latency, and context carryover across sessions.

When should I avoid using token compression on AI agent responses?▼

You should avoid token compression around destructive or irreversible actions, as the safety boundaries ensure exact error text and destructive-action prompts remain uncompressed.

Related Skills

Explore More Software Engineering Skills→
codenamevcodenamev
74

pragmatic-guard

Configure pragmatic mode with intensity levels, triggers, exemptions, and deferrals.

Community
Basic
SyntaxAsSpiralSyntaxAsSpiral
5

context-optimization

Compress, mask, and partition context to optimize AI agent token usage.

Community
Basic
UnitOneAIUnitOneAI
44

agentic-top-10

Audits agentic AI systems for OWASP Top 10 security risks across architectures.

Official
Advanced
jshsakurajshsakura
22

prompt-engineer

Refine prompts to enforce task boundaries, output contracts, and safety constraints.

Community
Advanced