Can I merge multiple research sessions into one Praetorian record?

Yes. Praetorian encourages reusing titles to merge new insights into existing compactions, preventing duplicate records and keeping your project knowledge base consolidated.

How does Praetorian differ from standard Claude memory?

Standard Claude memory is typically session-bound; Praetorian provides structured, distilled 'compactions' that persist across sessions and focus on high-value insights rather than raw transcripts.

What are 'nudges' in the Praetorian plugin?

Nudges are lightweight, one-line prompts (approx. 20 tokens) that appear at natural decision points, suggesting when to restore existing context or save new findings to keep your workflow efficient.

Will Praetorian increase my token usage?

No, it is specifically designed to reduce it. By restoring summarized insights instead of re-performing research or web searches, you save significantly on total context and tool-call costs.

Praetorian Memory & Context Manager

Name: Praetorian Memory & Context Manager
Author: Vvkmnn

byVvkmnn

0•

生産性とワークフロー

Manages cross-session memory by saving distilled insights and restoring context to reduce token waste and redundant research.

Praetorian acts as a sophisticated context guard for Claude Code, providing persistent cross-session memory through a system of intelligent nudges and distilled "compactions." By hooking into tool usage and subagent completions, it reminds users to restore existing knowledge before starting tasks and to save key findings once work is complete. This proactive management prevents token waste caused by re-exploring the same codebase or documentation, ensuring that hard-won insights and architectural decisions are preserved and easily accessible across multiple sessions without bloating the conversation history with raw, redundant data.

主な機能

01Seamless integration with Historian and Vigil for a comprehensive project state

02Token-efficient 'Summary First' restoration to verify relevance before full data loading

030 GitHub stars

04Automated workflow nudges triggered by tool usage and subagent completion

05Cross-session persistence that survives session resets and environment clears

06Intelligent 'compactions' that distill complex research into searchable, reusable insights

ユースケース

01Reducing token consumption in large projects by checking previously stored findings at session start

02Resuming complex research tasks across multiple days without re-running expensive web searches

03Capturing architectural decisions and rationale made during subagent explorations for future reference

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add vvkmnn/claude-emporium-beta claude-praetorian

For use in Claude.ai and ChatGPT

主な機能

01Seamless integration with Historian and Vigil for a comprehensive project state

02Token-efficient 'Summary First' restoration to verify relevance before full data loading

030 GitHub stars

04Automated workflow nudges triggered by tool usage and subagent completion

05Cross-session persistence that survives session resets and environment clears

06Intelligent 'compactions' that distill complex research into searchable, reusable insights

ユースケース

01Reducing token consumption in large projects by checking previously stored findings at session start

02Resuming complex research tasks across multiple days without re-running expensive web searches

03Capturing architectural decisions and rationale made during subagent explorations for future reference