Does it work automatically?

Yes, it uses triggers like SessionStart and PreCompact to automatically prompt you to save or restore context at the most effective moments.

Is an MCP server required?

Yes, this skill functions as a wrapper for the claude-praetorian-mcp server, which must be installed and configured in your Claude settings.

Can I search through previous sessions?

Yes, the skill includes a /search command that allows you to query past compactions to find specific decisions, patterns, or research.

How does Claude Praetorian save tokens?

It compacts detailed research and session history into concise summaries that are much smaller than raw chat history, saving 50-60% of tokens during restoration.

What kind of information is preserved?

It preserves web research, task results, codebase patterns, architectural insights, and technical rationales recorded during your sessions.

Claude Praetorian

Name: Claude Praetorian
Author: Ven0m0

byVen0m0

•

Productividad y Flujo de Trabajo

Optimizes and preserves session context automatically to reduce token usage and maintain project continuity across coding sessions.

Claude Praetorian is a specialized context management skill designed to bridge the gap between AI development sessions by automatically capturing and restoring critical project insights. It monitors session activity—including web research, codebase exploration, and architectural decisions—and prompts the user to compact these findings into persistent storage. By restoring these 'compactions' in subsequent sessions, it saves 50-60% on token costs while ensuring the AI retains deep knowledge of previous work without requiring redundant re-exploration of the codebase or documentation.

Características Principales

01Automatic session start restoration and end-of-session compaction prompts

02Project-specific detection to prioritize relevant past compactions automatically

03Significant token savings (50-60%) by loading compact insights instead of raw history

04Manual commands for manual compaction, restoration, and historical searching

05Context preservation for web research, subagent findings, and API documentation

061 GitHub stars

Casos de Uso

01Resuming complex coding tasks without the AI needing to re-read the entire codebase

02Reducing API costs by minimizing redundant context processing in large-scale projects

03Persisting deep research findings and technical decisions across multiple chat sessions

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add ven0m0/claude-config claude-praetorian

For use in Claude.ai and ChatGPT

Download Skill