Leaderboard/Prompting/directional-prompting
Last commit on May 21, 2026·Created on May 21, 2026

kingbootoshi/directional-prompting

Writes every prompt with a destination and a path forward.
Combined rank
#499
across all skills
In Prompting
#10
category rank
Stars
136
0.0% in last 7d
Forks
13
0.0% in last 7d
Watchers
0
0 watchers in last 7d
Traction scoreGitHub stars can be faked, so popularity alone can be misleading. Traction Score looks for broader signs of real attention, adoption, and active maintenance.
Description

This skill guides the writing of any prompt, system instruction, or skill description using a two-layer method: outcome-first to define the goal, success criteria, and stopping condition, and directional language to keep every sentence naming the correct action. It is aimed at prompt engineers, agent developers, and skill authors who need their instructions to be unambiguous and resistant to drift across turns. An included audit pass helps replace negative phrasing with positive replacements, leaving only the rare cases where negation is the cleaner choice.

WHO IT'S FOR
Prompt Engineers
write and audit system prompts
AI Agent Developers
craft agent directives and system instructions
AI Skill Authors
author skill descriptions and SKILL.md bodies
AI Safety and Alignment Researchers
design eval rubrics and instruction sets
30 Days of GitHub Stars
+22 stars
↑ 20.4% monthly change
Jul 7Jul 14Jul 21Jul 28Aug 5
Compatibility & Install
1 Skill · 5 Compatible agents · 2 categories · 2 install paths
Compatible agents
Claude CodeCodexCursorGemini CLIAntigravity
Categories
PromptingWriting
Install via
Clone to Claude
Repository instructions
Clone to Codex
Repository instructions
Also in this category
4 of 13