fable-prompting

Routes Fable prompting tasks into house-rules setup, goal loops, grading, and prompt drafting branches.

2|Updated Apr 15, 2026
One-click install
npx skills add https://github.com/tommylower/cortex --skill fable-prompting-tommylower
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: fable-prompting
Source: https://github.com/tommylower/cortex/tree/main/agent-workflows/fable-prompting
Command: npx skills add https://github.com/tommylower/cortex --skill fable-prompting-tommylower

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Prompting Fable like a current-generation model produces current-generation results. This Skill fixes that by routing you to the right workflow: installing house rules in a project, starting a self-checking loop on a goal, grading an artifact with a fresh sub-agent, or drafting a Fable prompt built on the goal-not-steps method. ## Core Features & Use Cases - Job Router: On invocation, asks which job you want (or detects it from your request) and runs that branch to its completion criterion. - House Rules Block: Drop-in CLAUDE.md guardrails covering invariants, a concrete bar for done, separate verification, and autonomy budgets. - Loop Template: A fill-in template for running Fable under /goal or /loop against a checkable bar, with a fresh sub-agent verifier so the builder never grades itself. - Use Case: You want Fable to rebuild a landing page until it matches a reference screenshot. Use the loop branch to fill GOAL, THE BAR, and FUEL, then run it under /loop while a separate grader sub-agent hunts for failures each pass. ## Quick Start Ask the agent to start a Fable loop on your goal with a concrete checkable bar using the fable-prompting skill.

Frequently Asked Questions about fable-prompting

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

FAQPage Schema
How do I prompt Fable to get better results?▼

Give Fable the goal instead of step-by-step instructions, fence it with house rules, and define a concrete self-checkable bar for done. Then loop it against that bar with a separate sub-agent verifying the real output.

How do I set up house rules for Fable in a project?▼

Use the house rules branch to paste the drop-in block into the project's CLAUDE.md and fill the bracketed invariants, key locations, and spend cap. Fable reads these guardrails every session so they are set once.

Should I use /goal or /loop for a Fable task?▼

Run under /goal with a turn cap when the bar fits in one sentence an evaluator can check; otherwise use /loop. The designing-loops skill covers choosing between loop primitives in more detail.

Why should the builder never grade its own work?▼

The building session is biased and carries a trajectory of reasons it is finished. Verification uses a fresh Fable sub-agent with separate context, pointed at the real artifact and told to prove it is failing.

When is ultracode worth the cost with Fable?▼

Almost never; a good loop with an ambitious goal usually suffices. The exception is foundations, a from-scratch system you will build on for months, where a strong base pays off long-term.