Tool-specific resourceFramework · 8 min

Which work deserves your best model?

A discovery prompt that inspects your real work and hands you ready-to-run briefs for the few tasks worth escalating to Claude Fable 5.

Use this the moment Fable is back, so you point your most expensive model at the work that actually deserves it, not whatever happens to be in front of you.

Free · open resource

Read it here, or download the PDF.

No email required. Copy the prompt, save this page, or download the PDF to keep.

Download the PDF

the idea

The skill is knowing which work to hand it.

Most people point their most powerful model at whatever is in front of them. That is backwards. The patient, expensive models earn their keep on one kind of work: big, spread across several tools, heavy on judgment, with a clear finish line. Short and obvious work belongs on a faster model or with you.

Escalate work that is large enough to justify a slower, more expensive run.
Down-rank work that is short, obvious, or highly interactive.
Keep work that must stay human out of the queue entirely.
The prompt does the sorting so you are not guessing.

how to use it

Point it at your real context, then choose.

Paste the prompt into Claude, fill the three bracketed sections with your role, the sources it can inspect, and your constraints, then run it. You get a ranked shortlist and ready-to-run briefs for the top three. It does not start the work. It finds the work worth doing.

Fill your role, priorities, and constraints.
List the sources it may inspect: repos, docs, backlogs, calendar, analytics.
Run it and read the ranked shortlist.
Pick one brief and hand it over.

bonus

9 prompts to run the second Fable is back.

While you wait, prepare every prompt you will send the moment it returns. Here are nine worth stealing. The full discovery prompt above is what finds the work worth escalating in the first place.

Review all code written after the date Fable was banned. Look for optimizations and improvements.
Walk every major user path in your app with browser control. Report where users get confused.
Make a checklist of every task you do until tonight. Ask Fable what it could automate.
Feed it your goals, skills, and assets. Ask what simple business could earn your first dollar online.
Connect the X MCP and find 5 use cases others run Fable for that fit your workflows.
Have it read your last 100 posts and pitch 5 SaaS ideas you could build.
/loop it every 24 hours to run a security check on all your API endpoints.
Use the Unreal 5.8 MCP to build in-depth 3D games.
Ask Sonnet 5 what incredible prompts you could give Fable tonight, based on what it knows about you.

Copy/paste prompt

Use this with your AI assistant

You are helping me decide which parts of my work are worth escalating to Claude Fable 5. Do not execute any of the work yet. Your job is to inspect my context, find strong candidates, and turn the best ones into ready-to-run Fable briefs.

My role and current goals:

[Describe your role, priorities, team, business goals, creative goals, or personal operating constraints.]

Use this context:

[List the sources you can inspect: repositories, project folders, docs, Notion, Slack, Linear/Jira, email, calendar, analytics, customer research, previous agent sessions, task lists, or anything else relevant.]

Constraints:

[List deadlines, privacy boundaries, accounts or tools you may not use, budget, sensitive areas, approval rules, and work that must remain human.]

First, inspect the available context. Do not brainstorm generic examples. Build an inventory of: active projects, repeated workflows, stalled decisions, messy backlogs, work that spans several tools, and work where better planning, judgment, verification, or follow-through would materially improve the result.

Then score each candidate 1 to 5 on: multi-source context, delegation fit, judgment required, clear finish line, leverage, and Fable fit. Recommend Fable only when the task is large enough to justify a slower, more expensive run. Down-rank work that is short, obvious, highly interactive, hard to verify, or better handled by a faster model or a human.

Return: the top 10 ranked use cases, the evidence for each, why each is or is not worth Fable, the expected deliverable, the verification method, the context and permissions needed, and the risks or human decisions required. Then write a ready-to-run brief for the top three (problem, outcome, sources, constraints, workflow, human checkpoints, evidence required). Stop after the list and the three briefs. Do not start any Fable task until I choose one.

Inside the resource

Use this when the idea is useful, but the next step is unclear.