Where is the log data stored?

Data is stored locally within your project in the .claude/RUNNING_LOG.md file, ensuring your process memory stays with your codebase.

Can this skill detect when Claude is making assumptions?

Yes, it uses advanced pattern-matching to identify when the AI expresses uncertainty or makes assumptions, logging them automatically for later validation.

How does the librarian function work?

The /review-backlog command analyzes your entries to suggest links between related items, harmonize tags, and help you prioritize your captured ideas.

How does the /idea command differ from standard logging?

The /idea command is designed for zero-friction capture, allowing you to record thoughts instantly while the AI automatically handles metadata like timestamps and tags.

Running Log

Name: Running Log
Author: jcmrs

byjcmrs

0•

Productividad y Flujo de Trabajo

Maintains a persistent, schema-driven log of ideas, AI-detected reasoning patterns, and external consultations across development sessions.

Running Log is a specialized process memory tool designed to bridge the gap between sessions by capturing developer ideas and AI reasoning patterns in a structured format. It utilizes a three-component architecture: an ultra-minimal human capture workflow via the /idea command, automated AI detection of assumptions and decisions, and a dedicated librarian function for organizing the backlog. By tracking consultations, uncertainty, and critical signals in a searchable Markdown format, it ensures project institutional knowledge is preserved and easily accessible, preventing context loss during complex, long-term development tasks.

Características Principales

01Profile-based configuration tailored for developers, researchers, and engineers.

02Comprehensive /review-backlog command for prioritizing and linking related entries.

03Ultra-minimal /idea command for rapid thought capture without interrupting workflow.

04Automated AI detection of assumptions, uncertainties, and critical decision forks.

05Intelligent tag harmonization and deduplication using Levenshtein similarity.

060 GitHub stars

Casos de Uso

01Auditing AI reasoning patterns to identify low-confidence areas or hidden project risks.

02Managing a lightweight backlog of feature ideas and notes directly from the CLI.

03Tracking architectural decisions and rejected alternatives across multiple coding sessions.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jcmrs/jcmrs-plugins running-log

For use in Claude.ai and ChatGPT