Workflows

Prompt chaining

Break a big task into linked steps where each output feeds the next.

Knowledge cards
Workflow
Add a Check Step Between the Links
Inserting an explicit verification step between the stages of a prompt chain: before a step's output becomes the next step's input, check it against clear criteria — and only proceed if it passes. It's a quality gate at each handoff rather than trusting every link to be perfect.
L3 · Workflows
Workflow
Fan Out Independent Work, Then Merge
A chaining shape for when a task divides into independent parts: instead of one long sequential chain, run each independent sub-task on its own — each in a fresh, focused context — then use a final step to synthesise the results into one coherent output. Fan out to do the parts, fan in to combine them.
L3 · Workflows
Workflow
Pass Structured Handoffs Between Chain Steps
A reliability technique for prompt chains: have each step emit its result in a defined structure — a labelled list, a table, JSON — that the next step consumes, instead of piping loose prose from one prompt into the next and hoping it parses.
L3 · Workflows
Workflow
Split a Task at Its Natural Seams, Not Arbitrarily
A chaining principle: break a large task into steps at its natural seams — points where the job genuinely changes, like extract → analyse → draft → refine — so each prompt does one thing well and hands a clean result to the next.
L3 · Workflows
Prompt chaining with Claude

One big task, several linked steps

Prompt chaining breaks a complex task into a sequence where each step's output feeds the next, instead of overloading a single prompt. Extract, then transform, then format; or research, then outline, then draft. Each step stays focused, which makes the output more reliable and the whole process easier to debug — when something goes wrong, you can see exactly which step produced it. A chain also lets you check and correct at each hand-off, so errors do not silently compound down the line.

Design the hand-offs

The craft in chaining is the interface between steps: what each step outputs and what the next expects. Ask Claude to produce structured output at each stage — a clear shape the next step can consume — rather than free prose you have to re-parse. Keep each step's context to what it actually needs, so the chain stays clean. Done well, a prompt chain turns an unreliable one-shot into a dependable little pipeline you can run again and trust.

Frequently asked
What is prompt chaining?

Breaking a complex task into a sequence of steps where each output feeds the next, rather than cramming everything into one prompt. Each step stays focused and the process is easier to debug.

Why use a prompt chain instead of one prompt?

Because focused steps produce more reliable output and let you check at each hand-off, so errors do not silently compound. You can also see exactly which step went wrong.

How do I design a good prompt chain?

Define the interface between steps — ask Claude for structured output the next step can consume — and keep each step's context to what it needs. Clean hand-offs make the chain dependable.