compression

Name: compression
Availability: InStock
Author: jacob-balslev

Select compression algorithms for SaaS payloads, Postgres storage, and agent-session contexts.

1|Updated May 6, 2026

Search Tags:#compression #context window #postgres #zstd #brotli #token pruning #semantic summarization

Ready to run

npx skills add https://github.com/jacob-balslev/skill-graph --skill compression-jacob-balslev

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: compression
Source: https://github.com/jacob-balslev/skill-graph/tree/main/marketplace/skills/compression
Command: npx skills add https://github.com/jacob-balslev/skill-graph --skill compression-jacob-balslev

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill prevents wasted bandwidth, storage, and context-window space by guiding correct compression choices for both infrastructure payloads and long-running agent sessions.

Core Features & Use Cases

SaaS payload compression guidance: Selects the right algorithm and settings for static assets and dynamic API JSON payloads, including content negotiation and cache safety.
Database storage compression strategy: Recommends Postgres TOAST Zstd usage and when to use application-layer compression for large blobs.
AI context compression techniques: Implements semantic summarization and token pruning workflows to remove dead context while preserving evidence paths and state.

Quick Start

Ask an AI agent to “compress my SaaS responses using zstd where appropriate, ensure Vary: Accept-Encoding is set, and compact the session context by summarizing intent/outcomes/evidence paths before continuing.”

Frequently Asked Questions about compression

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

FAQPage Schema

How do I reduce Postgres storage using zstd compression for large blobs?▼

Postgres storage compression uses TOAST Zstd to shrink large blobs and reduce disk space. Application-layer compression can also be applied to large blobs when database-level strategies are insufficient for your payload types.

What is the best way to compress dynamic API JSON payloads for SaaS applications?▼

Compressing dynamic API JSON payloads requires selecting the correct algorithm and settings for static assets and dynamic responses. Proper content negotiation and cache safety must be maintained by setting Vary: Accept-Encoding during SaaS payload optimization.

How does semantic summarization and token pruning work for AI agent sessions?▼

Semantic summarization and token pruning compact AI agent sessions by removing dead context. This context compression preserves evidence paths and state, preventing wasted tokens in the context window across long-running tasks.

When should I use application-layer compression instead of Postgres TOAST?▼

Application-layer compression is recommended over Postgres TOAST Zstd when handling specific large blobs that require custom compression strategies. This approach ensures correct compression choices are applied to the right data types and workflows.

Does SaaS payload compression require Accept-Encoding negotiation?▼

Yes, SaaS payload compression requires Accept-Encoding negotiation with a specific priority. This content negotiation ensures correct algorithm selection and maintains cache safety for dynamic API JSON payloads and static assets.