run-plan

Discover work items and propose run groupings from state.yaml.

201|19|Updated Dec 24, 2025
One-click install
npx skills add https://github.com/fabriqaai/specs.md --skill run-plan-fabriqaai
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: run-plan
Source: https://github.com/fabriqaai/specs.md/tree/main/src/flows/fire/agents/builder/skills/run-plan
Command: npx skills add https://github.com/fabriqaai/specs.md --skill run-plan-fabriqaai

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill helps planning by discovering available work items and suggesting optimal groupings before a run executes, reducing manual scoping and rework.

Core Features & Use Cases

  • Discover pending work items across intents and their dependencies.
  • Propose execution options: one-at-a-time, grouped by mode, or all-at-once.
  • Respect dependencies and user preferences, producing concrete run plans and enabling efficient handoffs to run-execute.
  • Integrate with state.yaml to reconcile items, track history, and record planning decisions.

Quick Start

After work-item decomposition, run this skill to receive recommended run scope and item groupings. Review the suggested option, adjust if needed, and trigger run-execute with the selected work items.

Frequently Asked Questions about run-plan

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

FAQPage Schema
How do I plan run scope for multiple work items with dependencies?

To plan run scope, the skill discovers available work items across intents, analyzes their dependencies, and suggests optimal groupings to run items single, batched, or all at once.

How does filesystem discovery reconcile pending work items with state.yaml?

Filesystem discovery reconciles with state.yaml by scanning available work items across intents and cross-checking them against the recorded state to track history and output concrete run options.

Can I group work items by execution mode for batched runs?

Yes, you can group work items by execution mode. The skill proposes multiple execution options, allowing you to batch items by mode, run them one-at-a-time, or execute all at once.

What is the best way to determine whether to run items single, batched, or all at once?

The best way to determine run execution is by applying the skill's planning logic, which evaluates work item dependencies and user preferences to recommend whether to run single, batched, or all at once.

Does the run-planning skill record user preferences for future scoping?

Yes, the run-planning skill records user preferences. It integrates with state.yaml to reconcile items, track planning decisions, and save preferences to optimize future run scope suggestions.

When should I not use automated work item grouping for run scope?

You should not use automated work item grouping for single-item workloads, as the skill's dependency analysis and batched execution options are specifically designed for multi-item workloads across intents.