Does Arc-Reflecting automatically modify my project rules?

No. While it identifies rule violations, it is non-prescriptive. It reports violations to the user and asks for permission before any updates are made to CLAUDE.md.

Where are the processed insights and instincts stored?

Reflections and patterns are saved locally to your home directory at ~/.claude/diaryed/, organized into 'global' and 'project-specific' subdirectories.

What is the difference between /reflect and /recall?

/reflect is used to analyze past diaries and save new patterns into 'instincts,' whereas /recall is used to retrieve those previously learned patterns during an active session.

How many diaries are required to identify a pattern?

The skill requires a behavior to appear in at least 3 diary entries to be classified as a 'Pattern.' Observations appearing only 1-2 times are noted but not promoted to pattern status.

Arc Reflecting: AI Session Insights

Name: Arc Reflecting: AI Session Insights
Author: GregoryHo

byGregoryHo

0•

生产力与工作流

Analyzes coding diaries to extract recurring patterns, user preferences, and rule violations across multiple AI development sessions.

Arc Reflecting is a specialized workflow skill designed to bridge the gap between individual AI coding sessions and long-term behavioral adaptation. By scanning session diaries, the skill identifies repeated decisions, architectural preferences, and violations of established project rules (CLAUDE.md). It employs a strict evidence-based approach, promoting observations to 'Patterns' only after three occurrences, and saves these insights as 'instincts' that Claude can recall in future sessions. This allows the AI to develop a deep understanding of a developer's unique style and project requirements over time.

主要功能

01Cross-project and global preference tracking

02CLAUDE.md rule violation detection with evidence

03Evidence-based 'instinct' generation and saving

040 GitHub stars

05Automated pattern extraction from session diaries

06Multi-strategy diary scanning (unprocessed, project-focused, or recent window)

使用场景

01Detecting and correcting persistent AI rule violations in large-scale projects

02Summarizing developer habits and session learnings for long-term workflow optimization

03Identifying recurring architectural preferences across multiple microservices

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add gregoryho/arcforge arc-reflecting

For use in Claude.ai and ChatGPT

Download Skill