handcraft

Enforce one-change-per-turn software development with approval-gated scoped tooling.

1|Updated May 28, 2026
One-click install
npx skills add https://github.com/SilvrDuck/skills --skill handcraft
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: handcraft
Source: https://github.com/SilvrDuck/skills/tree/main/skills/handcraft
Command: npx skills add https://github.com/SilvrDuck/skills --skill handcraft

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

It prevents an AI from running ahead, making bundled assumptions, or “doing extra” by enforcing a strict one-change-per-turn workflow that keeps the user firmly in control of architecture and decisions.

Core Features & Use Cases

  • One-change-per-turn build loop: Each turn produces the smallest meaningful diff, runs targeted tooling, and asks exactly one next-step question.
  • Approval-gated spec and surface mapping: Captures the user’s intent, infers the public interface (endpoints/modules/data shapes/commands), and only then moves to implementation.
  • Tooling-verification discipline: Runs LSP/formatters/tests scoped to touched files, distinguishes expected vs unexpected errors (e.g., stubs/forward-refs), and keeps unresolved imports from slipping through unintentionally.
  • External API call safety checks: Flags unverified library/API usage when docs access is unavailable or mismatch is detected when it is available.
  • Deferred-gap tracking: If the user defers decisions, the skill re-surfaces them at the moment they become load-bearing.

Quick Start

Activate handcraft mode by telling the agent to build your feature function by function, approving each step before it proceeds.

Frequently Asked Questions about handcraft

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

FAQPage Schema
How do I control an AI coding agent to build features function by function?

To control an AI coding agent building features function by function, use an approval-gated workflow that enforces a one-change-per-turn loop, producing minimal diffs and asking exactly one next-step question before proceeding.

How does an approval-gated coding workflow handle stubs and forward references?

An approval-gated coding workflow handles stubs and forward references by distinguishing expected errors from unexpected ones, verifying touched files with LSP and formatters, and keeping unresolved imports from slipping through unintentionally.

What is the best way to refactor code incrementally with deterministic diffs?

The best way to refactor code incrementally with deterministic diffs is to enforce a one-change-per-turn loop with scoped tooling verification, capturing user intent through a spec surface map before implementation.

How do I prevent an AI from making bundled assumptions during software development?

To prevent an AI from making bundled assumptions during software development, enforce a strict one-change-per-turn workflow that keeps the user in control of architecture, re-surfacing deferred decisions when they become load-bearing.

Do I need LSP and formatter pre-flight detection for iterative coding?

Yes, iterative coding requires pre-flight detection for guidelines, LSPs, formatters, and runners, which then scopes tooling verification to touched files and flags unverified external API calls when docs access is unavailable.

When should I not use a one-change-per-turn build loop?

You should not use a one-change-per-turn build loop when you need rapid prototyping or large bundled changes, as its strict approval-gated workflow is designed for deliberate, minimal diffs and controlled handling of deferred gaps.