fable-5-train-opus-skills-after-it-retires  by tomicz

AI skill library generation framework

Created 1 week ago

New!

337 stars

Top 81.4% on SourcePulse

GitHubView on GitHub
Project Summary

This project addresses critical knowledge transfer upon a senior engineer's retirement. It aims to create a comprehensive skill library for junior engineers and smaller AI models (Sonnet-class) to independently maintain, debug, extend, and advance the project, ensuring continuity and cost-efficiency.

How It Works

A multi-agent orchestration approach documents project knowledge and generates executable skills. Phase 1 involves deep repository investigation (README, build, tests, CI, git history). Phase 2 authors a structured library of 10-16 skills (CORE and ADVANCED) using an imperative, copy-pasteable runbook voice. Emphasis is on correctness and ground truth, ensuring skills are self-contained and usable by less experienced personnel or AI models.

Quick Start & Requirements

The README details a process for generating a skill library, not direct project installation. Target users are junior/mid-level engineers and Sonnet-class AI models. Prerequisites include the project's existing codebase and development environment, which are unspecified. The primary output is a set of skills within the .claude/skills/ directory.

Highlighted Details

  • Skill Taxonomy: Mandates 10-16 skills covering CORE (change control, debugging, architecture, config, build, ops, diagnostics, validation, docs) and ADVANCED (campaigns, analysis toolkits, research frontiers, methodology) functionalities.
  • Authoring Rules: Strict guidelines include imperative runbook voice, YAML frontmatter, grounding claims in repo facts, defining jargon, and self-contained knowledge embedding.
  • Knowledge Transfer: Prioritizes capturing implicit knowledge and operational discipline to mitigate the bus factor.
  • Cost-Conscious Design: Skills target "cheaper" models like Sonnet, optimizing execution efficiency.

Maintenance & Community

The project's context is knowledge transfer due to a key engineer's retirement. No information is provided regarding ongoing community engagement, external contributors, sponsorships, or future roadmaps beyond the skill library's creation.

Licensing & Compatibility

No licensing information is specified. Compatibility for commercial use or closed-source linking cannot be determined.

Limitations & Caveats

This repository acts as a prompt for skill library generation, not a deployable application. Success depends on thorough investigation and skill quality. The README lacks details on the project's codebase, build system, or dependencies, hindering assessment of its current operational status. Effectiveness relies on the author's ability to answer specific questions about project challenges and history.

Health Check
Last Commit

1 week ago

Responsiveness

Inactive

Pull Requests (30d)
1
Issues (30d)
1
Star History
336 stars in the last 9 days

Explore Similar Projects

Feedback? Help us improve.