skills.rest
matthewharwood
matthewharwoodCOMMUNITY
Explore all skills by this author →
GitHub

run-prompt

Execute saved Markdown prompt files sequentially or in parallel with state handoffs.

Updated May 23, 2026
Search Tags:#prompt library#agentic workflows#prompt execution#parallel runs#state handoff#CLI compatibility
One-click install
npx skills add https://github.com/matthewharwood/dean-n-dean --skill run-prompt-matthewharwood
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: run-prompt
Source: https://github.com/matthewharwood/dean-n-dean/tree/main/.agents/skills/run-prompt
Command: npx skills add https://github.com/matthewharwood/dean-n-dean --skill run-prompt-matthewharwood

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It solves the need to reliably run one or more previously saved prompt instructions from your local prompts directory without manually copying and re-executing them.

Core Features & Use Cases

  • Prompt discovery by convention: Loads prompts from ./prompts/ using a numeric filename pattern like 001-*.md.
  • Configurable execution: Supports sequential execution by default and parallel execution when multi-agent tooling is available.
  • Compatibility flags: Accepts legacy command syntax such as --model, --effort, and --auto and applies them as supported (with clear fallbacks when not supported).
  • State handoff for multi-step chains: Writes durable handoff notes to ./prompts/.state/[number]-handoff.md between steps where continuation depends on prior decisions.

Quick Start

Run the prompt chain by specifying the prompt numbers (for example, run 001 through 003 sequentially) so the skill loads each matching ./prompts/ file, applies effort precedence, and executes in the selected mode.

Frequently Asked Questions about run-prompt

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

FAQPage Schema
How do I execute saved prompt instructions from a directory for agentic workflows?▼

To execute saved prompts, this skill parses numeric IDs from arguments and loads matching Markdown files from the ./prompts/ directory. It runs prompts sequentially or in parallel for multi-step agentic workflows without manual re-execution.

What naming convention do prompt files need to follow for sequential execution?▼

Prompt files must use a numeric filename pattern like [number]-*.md in the ./prompts/ directory. The skill parses numeric prompt IDs from your arguments to discover and load the correct saved instructions for execution.

How does state handoff work when running multi-step prompt chains?▼

State handoff in multi-step chains writes durable notes to ./prompts/.state/[number]-handoff.md between steps. This ensures continuation decisions from prior steps are passed to dependent prompts during sequential execution.

Can I run multiple prompts in parallel mode for multi-agent tooling?▼

Running multiple prompts in parallel is supported when multi-agent tooling is available. By default, execution is sequential, but the skill enables parallel execution to optimize agentic workflows and multi-step development sessions.

Does this prompt execution tool support legacy CLI compatibility flags like --model and --effort?▼

Legacy command syntax such as --model, --effort, and --auto is accepted and applied as supported. The skill applies effort precedence from CLI flags, file hints, and defaults, providing clear fallbacks when flags are unsupported.

Related Skills

Explore More Software Engineering Skills→
79yuuki79yuuki

prompt-design

Analyze and restructure agent-facing prompts for clearer execution.

Community
Intermediate
auerbachbauerbachb
3

pm-handoff

Generate a clipboard-ready PM handoff prompt from repository config, GitHub state, and memory.

Community
Advanced
RobbyMo81RobbyMo81

prose

Coordinate multi-agent workflows in OpenProse with sessions, imports, and state backends for parallel execution and persistence.

Community
Advanced
a2musa2mus

speckit-implement

Execute task breakdowns to implement features and tests with TDD.

Community
Advanced