1marcelserrano/prompt-chain

prompt-chain

English · Português

Break complex multi-step work into self-contained prompts that survive context limits — chained (sequential) or fanned out (independent).

The whole skill is a single markdown file. No dependencies, no server, no external state store. Built for anyone running long multi-phase work in Claude — developers and non-developers alike.

Install • How it works • When to use • FAQ

---

A Claude skill for distributing work across fresh chat sessions, in two modes: CHAIN (each prompt generates the next, carrying full context forward, until done) and FAN-OUT (one independent prompt per piece, no ordering, run in any order). Cold-start safe. Context-limit immune.

How it works

Two modes, picked by one question — do the pieces depend on each other's output or state?

CHAIN (yes — order is load-bearing): each prompt runs one stage and emits the next.

flowchart LR
    A[You paste<br>STAGE 1] --> B[Chat 1 executes<br>+ emits STAGE 2 prompt]
    B -->|copy-paste| C[Chat 2 executes<br>+ emits STAGE 3 prompt]
    C -->|copy-paste| D[Chat N executes]
    D --> E[CHAIN COMPLETE<br>all deliverables listed]
    B -.blocker.-> P[CHAIN PAUSED<br>asks, never guesses]

FAN-OUT (no — pieces are independent): one dispatcher emits N self-contained prompts; each runs alone, in any order, sharing the same stable context.

flowchart LR
    A[You paste<br>DISPATCHER] --> B[emits N prompts<br>same shared context]
    B --> P1[Piece 1<br>own chat]
    B --> P2[Piece 2<br>own chat]
    B --> P3[Piece N<br>own chat]
    P2 -.blocker.-> Q[asks, never guesses<br>others unaffected]

| Without prompt-chain | With prompt-chain | |----------------------|-------------------| | One mega-prompt, context fills up, output degrades, you lose state when the chat dies. | A first prompt sets the work. Each chat finishes its piec