efficient-fable

Delegate token-intensive research, coding, and testing tasks to cheaper subagents.

Updated Apr 3, 2026
One-click install
npx skills add https://github.com/lapaixkemsdortshlee-svg/AyitiMarket --skill efficient-fable
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: efficient-fable
Source: https://github.com/lapaixkemsdortshlee-svg/AyitiMarket/tree/main/.agents/skills/efficient-fable
Command: npx skills add https://github.com/lapaixkemsdortshlee-svg/AyitiMarket --skill efficient-fable

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Efficient Fable helps you handle broad, codebase-heavy tasks without spending expensive model tokens on every scan, log pass, or repetitive edit. It keeps the strongest judgment focused on the hard decisions while routine exploration and testing are delegated.

Core Features & Use Cases

  • Breaks large tasks into research, coding, and testing lanes.
  • Uses cheaper subagents for repository scans, documentation summaries, bounded bug hunts, browser checks, and narrow code edits.
  • Keeps the orchestrator focused on architecture, product tradeoffs, conflict resolution, synthesis, and final review.
  • Best for unfamiliar repositories, long test output, multi-file changes, and validation work that can run in parallel.
  • Example: Ask it to investigate a failing feature across several files, gather targeted evidence, and return one coherent implementation plan.

Quick Start

Use efficient-fable to split a large repository task into parallel research, coding, and testing passes and return a judgment-ready summary with evidence.

Frequently Asked Questions about efficient-fable

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I reduce token waste when analyzing unfamiliar codebases?

Reduce token waste in unfamiliar codebases by delegating token-intensive repository scans, documentation summaries, and log passes to cheaper subagents. This keeps expensive model tokens focused on architecture, product tradeoffs, and final review rather than routine exploration.

What is the best way to coordinate parallel research, coding, and testing across multiple files?

The best way to coordinate parallel multi-file work is to split large tasks into research, coding, and testing lanes. Cheaper subagents gather targeted evidence and run bounded bug hunts, while the orchestrator synthesizes results into one coherent implementation plan.

Can I use subagent delegation for long test output and broad documentation scans?

Yes, you can delegate long test output and broad documentation scans to subagents. These routine validation workflows and repetitive edits run on cheaper models, returning evidence-rich handoffs with explicit reports of files, diffs, and failures for expert synthesis.

Does orchestrating codebase-heavy tasks require explicit verification commands?

Orchestrating codebase-heavy tasks requires explicit verification commands or browser flows. Bounded task scopes and strict stop conditions ensure subagents return verified evidence, allowing the orchestrator to resolve conflicts and perform final judgment safely.

When should I not use subagent delegation for repository work?

Avoid subagent delegation for repository work when tasks lack clear boundaries or require deep, continuous architectural judgment. Delegation works best for narrow code edits and parallel validation, not for unbounded exploration or complex product tradeoff decisions.

How do I generate a judgment-ready implementation plan from scattered repository evidence?

Generate a judgment-ready implementation plan by dispatching subagents for targeted evidence gathering across files. They return bounded reports of diffs and uncertainties, which the orchestrator synthesizes into a single coherent plan ready for expert review.