Home /Claude Skills /superwiser

Superwiser

Superwiser enhances Claude Code's personalization by capturing an...
Core Principle:
Superwiser is a plugin specifically designed for Claude Code to capture user preferences and correct instructions during AI coding sessions. It stores this information in a local searchable database for reference in future sessions, reducing redundant guidance. The project employs a lightweight architecture that analyzes only user prompts, ensuring efficiency and privacy while enhancing development productivity.
KEY FEATURES
01Rule Extraction
Automatically extracts coding rules from user prompts using Sonnet model for efficient background processing.
02Conflict Detection
Detects contradictory guidance across sessions and resolves rule conflicts via user prompts.
03Importance Scoring
Scores rules based on search hits and other factors, prioritizing high-weight preferences.
04Local Storage
Stores rules in SQLite database, ensuring data privacy with local retention.
github.com/mu4farooqi/superwiser
research·mu4farooqi·2026-01-14·5·🔱 0
Curated by agent-skills.cc
Installation
Download
HTTPS
git clone https://github.com/mu4farooqi/superwiser.git
SSH
git clone [email protected]:mu4farooqi/superwiser.git
GitHub CLI
gh repo clone mu4farooqi/superwiser
FAQ
Q: What are the installation steps for Superwiser Agent Skills?
1.User Prompt Submission: Input preferences or corrections in Claude session.
2.Hook Queue Extraction: Submit hook queues prompt for rule extraction.
3.Background Rule Processing: Background worker extracts and stores rules using Claude.
4.Future Search Reference: Stored rules are searchable by Claude in new sessions.
Q: What are the highlights of Superwiser Agent Skills?
  • Lightweight processing: Analyzes only user prompts, reducing token usage by over 90%.
  • Local data storage: SQLite database ensures privacy with no external service dependency.
  • Automatic conflict detection: Identifies contradictory rules promptly for user resolution.
  • Project context discovery: Periodically generates documents to help AI understand code architecture.
Q: What are the use cases for Superwiser Agent Skills?
  • Customizing AI behavior in development, e.g., specifying databases or error handling.
  • Loading historical preferences to optimize AI output accuracy in new sessions.
  • Importing preferences from past conversations to save configuration time.
  • Adjusting plugin settings via commands to fit varied project needs.
Q: What are the limitations of Superwiser Agent Skills?
  • Only compatible with Claude Code version 1.0.33 and above.
  • Depends on Python 3.10+ environment, potentially limiting deployment flexibility.