Skip to content

DigitalHumanitiesCraft/promptotyping-skill

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

18 Commits
 
 
 
 
 
 
 
 

Repository files navigation

Promptotyping

A context engineering methodology for LLM-assisted development of research artifacts (tools, visualizations, editions), packaged as an Agent Skill.

Promptotyping treats documents as the source of truth and code as disposable. Instead of prompting for code directly, you build compressed, domain-adaptive docs (knowledge, requirements, design, implementation) that encode your decision logic. The result: reproducible reasoning across sessions, agents, and team members. If the code is lost, the knowledge isn't.

The skill guides the agent through four phases (Preparation, Exploration, Distillation, Implementation) and includes five operations as reference files that the agent loads on demand.

Follows the open Agent Skills standard. Primary target: Claude Code. Compatible with any agent supporting the standard (Codex CLI, Cursor, Gemini CLI, GitHub Copilot, etc.).

Installation

# Via skills.sh package manager (auto-detects your agent)
npx skills add DigitalHumanitiesCraft/promptotyping-skill

# Or manually: user-level (all projects)
cp -r promptotyping ~/.claude/skills/    # Claude Code
cp -r promptotyping ~/.codex/skills/     # Codex CLI
cp -r promptotyping ~/.gemini/skills/    # Gemini CLI
cp -r promptotyping ~/.cursor/skills/    # Cursor

# Or manually: project-level
cp -r promptotyping .claude/skills/       # Claude Code
cp -r promptotyping .agents/skills/       # Codex CLI, Gemini CLI, Cursor

Usage

Ask the agent to use the methodology:

  • "Use promptotyping for this project"
  • "Promptotyping für dieses Projekt"

Operations during a session (Claude Code syntax, other agents may differ):

  • /promptotyping orient – detect project state, start session
  • /promptotyping handoff – commit, journal, end session
  • /promptotyping check – gap analysis, update docs
  • /promptotyping distill – create docs from exploration (setup)
  • /promptotyping verify – validate external facts via web search

Background

License

MIT

About

Agent Skill: Context Engineering methodology for LLM-assisted research artifact development (agentskills.io standard)

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors