What is the difference between takt-analyzer and takt-optimize?

Takt-analyzer is a read-only diagnostic tool that generates Markdown reports on issues and inefficiencies. Takt-optimize is an execution tool that directly modifies or generates files to apply those optimizations.

Which version of TAKT is supported?

This skill is optimized for TAKT v0.31.0 and later, including support for the v0.30.0 logging path changes and provider_options structures.

Does this skill require execution logs to work?

No, it performs static analysis of your YAML pieces and facets by default. If logs in .takt/logs/ are available, it will automatically include dynamic diagnostic data like loop hotspots and rule match statistics.

How do I trigger the analysis?

You can trigger it with natural language commands like 'analyze piece', 'takt analyze', 'check facet quality', or 'diagnose takt logs' within Claude Code.

TAKT Piece & Workflow Analyzer

Name: TAKT Piece & Workflow Analyzer
Author: j5ik2o

byj5ik2o

•

分析と監視

Analyzes TAKT pieces and facets to detect structural issues, style guide violations, and execution inefficiencies.

The TAKT Analyzer is a specialized diagnostic tool for developers using the TAKT framework to orchestrate AI agents. It performs deep static analysis of piece YAML structures and facet content to ensure consistency, style guide compliance, and optimal rule design. By ingesting execution logs (.jsonl), the skill provides dynamic insights into workflow performance, identifying loop hotspots, high abort rates, and rule evaluation bottlenecks. It serves as an essential quality assurance layer, helping users transition from custom, experimental configurations to robust, production-grade AI workflows.

主な機能

01Log-based diagnostic analysis of rule evaluation efficiency and abort rates

02Structural YAML validation for piece movements and transition logic

03Automated identification of custom facets that can be replaced by built-in alternatives

04Detection of responsibility overlaps and facet separation violations

05Facet quality assessment against persona, policy, and instruction style guides

0614 GitHub stars

ユースケース

01Optimizing API costs by identifying and reducing AI-fallback rule evaluations

02Debugging complex state transitions and loop monitors in TAKT pieces

03Performing health checks on large-scale AI agent workflows using execution logs

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add j5ik2o/ai-tools takt-analyzer

For use in Claude.ai and ChatGPT