AgusRdz

Chop — DevOps skill for Claude Code

DevOps community

CLI output compressor for Claude Code.

How to install Chop

This entry records only its repository, not the path inside it, so there is no exact command to give. Open AgusRdz/chop and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Chop does

CLI output compressor for Claude Code. Reduces token consumption by 50–90% by compressing verbose command output before it enters the context window. Supports 52+ commands — git, docker, kubectl, npm, terraform, and more.

Alternatives in DevOps

Kubeshark — Cluster-wide network observability for Kubernetes 11.8k ★
Codeg — Aggregate and browse AI coding agent sessions (Claude Code, Codex, Gemini CLI, etc.) in one place 850 ★
Cloudflare Skill — Comprehensive Cloudflare platform reference docs for AI/LLM consumption 709 ★

README

chop

**CLI output compressor for Claude Code, Gemini CLI, Codex CLI, and Antigravity IDE.**

Claude Code and other AI agents waste 50-90% of their context window on verbose CLI output — build logs, test results, container listings, git diffs. **chop** compresses that output before Claude sees it, saving tokens and keeping conversations focused.

The name comes from _chop chop_: the sound of something eating through all that verbosity before it ever reaches the context window.

How It Works

When Claude Code runs a Bash command, the raw output is fed back into the conversation as a `tool_result` — part of the **input** of the next API call. `chop` intercepts that result and compresses it before it enters the context.

sequenceDiagram
    participant CC as Claude Code
    participant H as PreToolUse Hook
    participant CH as chop
    participant API as Claude API

    CC->>H: bash("docker ps")
    H->>CH: raw output (850 tokens)
    CH-->>H: compressed output (250 tokens)
    H->>API: tool_result (250 tokens)
    API-->>CC: response

The Cascade Effect

The savings aren't one-time. Every `tool_result` that enters the context **stays there** for the rest of the session — included in the input of every subsequent API call. A bloated command result compounds across turns.

graph TD
    T1["Turn 1: docker ps → 850 tokens added"]
    T2["Turn 2: +850 tokens carried forward"]
    T3["Turn 3: +850 tokens carried forward"]
    TN["Turn N: +850 tokens carried forward"]
    COMPACT["Compaction triggered early ⚠️"]

    T1 --> T2 --> T3 --> TN --> COMPACT

    style COMPACT fill:#f66,color:#fff

graph TD
    T1C["Turn 1: chop docker ps → 250 tokens added"]
    T2C["Turn 2: +250 tokens carried forward"]
    T3C["Turn 3: +250 tokens carried forward"]
    TNC["Turn N: +250 tokens carried forward"]
    SESSION["Longer session, more context budget