skinny.cloud

cloud provides a script-based auditing tool designed to analyze Claude Code runtime environments for unnecessary context and memory usage. It addresses the issue of excessive resource consumption in AI operations by identifying what is loaded into the Claude Code setup before any prompt is entered, highlighting both visible and hidden memory costs. The tool is intended for those managing or optimizing Claude Code projects and seeking to reduce inefficiencies that can lead to higher operational costs.

The process begins with a single shell command that requires no installation or user account. Users paste the command into their terminal, after which the script automatically fetches, runs, and inspects the runtime environment. md files, hooks, and memory surfaces, providing immediate feedback in the form of a "Runtime Diet Card" displayed directly in the terminal. This card details the largest visible and hidden memory surfaces, identifies hooks, servers, skills, and agents in use, and points out the top suspect for excessive context usage, such as long-lived session transcripts or old project branches.

cloud generates a set of actionable "next session rules" based on the findings from the user's actual setup. These rules are intended to guide users in reducing unnecessary context for future sessions, such as starting with a clean slate or pruning outdated project directories. Additionally, a comprehensive report folder is created in the user's project directory containing the diet card, worksheet, baseline, and other supporting files to facilitate ongoing optimization efforts.

The tool is free to use and does not require any signup, email, or dashboard access. It is delivered as a shell script that can be run directly or cloned from a public repository. cloud is positioned as a solution for lean AI operations, providing transparency into context and memory usage within Claude Code environments.