Prompt Chaining for Complex Tasks

# Prompt Chaining for Complex Tasks

You are an expert in prompt pipeline design. Break a complex task into a chain of focused, single-purpose prompts where each output feeds the next — a technique that consistently outperforms single-prompt approaches on sophisticated tasks.

## Complex Task

**Task that resists a single prompt:** [COMPLEX_TASK]
**Why a single prompt fails here:** [REASON — e.g., too many competing objectives, early steps constrain later ones, quality degrades when steps are combined]

## Chain Design

Break this task into [CHAIN_LENGTH] sequential prompt steps. For each step:

**Step 1 — [STEP_1_NAME]**
Purpose: [STEP_1_PURPOSE]
Input to this step: [INPUT]
Prompt: [STEP_1_PROMPT]
Expected output: [STEP_1_OUTPUT]
Quality check before proceeding: [CHECKPOINT]

**Step 2 — [STEP_2_NAME]**
Purpose: [STEP_2_PURPOSE]
Input: Output from Step 1 plus [ADDITIONAL_INPUT]
Prompt: [STEP_2_PROMPT]
Expected output: [STEP_2_OUTPUT]
Quality check: [CHECKPOINT]

[Continue for all [CHAIN_LENGTH] steps]

## Handoff Protocol

When passing output from one step to the next, use this handoff template:
"The previous step produced: [OUTPUT]. Use this as the foundation for the next step. Do not revisit or modify the previous output — build forward from it."

## Quality Gate

Define the minimum quality standard each step must meet before the chain continues. If a step fails the gate: [FAILURE_PROTOCOL].