preflight

Produces a structured preflight assessment of unknowns, risks, tensions and readiness for a given task.

26|3|Updated Oct 9, 2025
One-click install
npx skills add https://github.com/davidlee/spec-driver --skill preflight-davidlee
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: preflight
Source: https://github.com/davidlee/spec-driver/tree/main/.spec-driver/skills/preflight
Command: npx skills add https://github.com/davidlee/spec-driver --skill preflight-davidlee

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Preflight prevents wasted work by bounding up-front research, surfacing unknowns and tensions, and assessing readiness without drifting into implementation or open-ended exploration.

Core Features & Use Cases

  • Bounded research setup: Establishes what must be learned next and when to stop searching to preserve tokens and focus.
  • Readiness assessment with guardrails: Produces a structured critical assessment (confirmed inputs, assumptions, unresolved questions/risks, tensions/ambiguities) before declaring readiness.
  • Scope-bound decision support: Confirms that routing has already selected the governing skill path and that preflight does not replace that routing decision.

Quick Start

Use /preflight after routing has already established the next step requires bounded research, and ask it to surface remaining unknowns and explicitly define stopping conditions for the task: $ARGUMENTS.

Frequently Asked Questions about preflight

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

FAQPage Schema
How do I surface unknowns and risks before starting software implementation?

To surface unknowns and risks before implementation, use bounded up-front research to separate resolved facts from conscious assumptions and assess readiness constraints. This structured readiness assessment prevents wasted work by defining stopping conditions before substantive investigation begins.

What is the best way to bound research and preserve tokens during spec-driven development?

The best way to preserve tokens during spec-driven development is establishing bounded research setup that defines what must be learned next and when to stop searching. This prevents drifting into open-ended exploration and maintains focus on specific readiness criteria.

How do I create a readiness assessment with confirmed inputs and unresolved questions?

To create a readiness assessment with confirmed inputs and unresolved questions, structure a critical evaluation that separates resolved facts, conscious assumptions, and tensions or ambiguities. This artifact-forward analysis explicitly defines stopping conditions before declaring readiness for the next phase.

When do I need bounded research instead of open-ended exploration for design documentation?

You need bounded research for design documentation when routing has established the governing skill path and the next step requires substantive investigation rather than coding. It satisfies stop-conditions and token-efficiency without replacing the initial routing decision or drifting into implementation.

Can I use preflight to replace routing decisions for spec-driven development?

No, you cannot use preflight to replace routing decisions. Preflight operates under scope-bound decision support guardrails, confirming that routing has already selected the governing skill path and ensuring the bounded research phase does not override that established routing decision.