goal-grill

Elicit Vision, Outcome, and Eval criteria before implementation.

1|Updated Mar 22, 2026
One-click install
npx skills add https://github.com/COCOSiL-inc/COCOSiL_Analysis_V2 --skill goal-grill
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: goal-grill
Source: https://github.com/COCOSiL-inc/COCOSiL_Analysis_V2/tree/main/.claude/skills/goal-grill
Command: npx skills add https://github.com/COCOSiL-inc/COCOSiL_Analysis_V2 --skill goal-grill

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

goal-grill helps you remove uncertainty before you build by turning a vague task into a clear Vision, measurable Outcome, and concrete Eval/acceptance criteria.

Core Features & Use Cases

  • Vision / Outcome / Eval (3-layer goal tightening): Ask only the minimum questions needed to define what you’re trying to achieve and how you’ll judge success.
  • Session-based workflow (NEW / RESUME / REVIEW): Continue from a saved goal-grill session, or refine an existing goal document.
  • Writeout gate with explicit approval: Drafts a preview and waits for your approval before writing to docs/output/goals/<slug>.md.
  • Optional web-escalation for facts/versions/specs: Requests web checks only when needed to fill in uncertain, time-sensitive, or spec-driven parts of the goal.

Quick Start

Ask an AI assistant to run the goal-grill skill to define the vision, outcome metrics, and acceptance criteria for your new feature before you write any code.

Frequently Asked Questions about goal-grill

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

FAQPage Schema
How do I clarify undefined goals and acceptance criteria before starting implementation?

To clarify undefined goals and acceptance criteria, you can use a 3-layer goal tightening workflow that elicits Vision, Outcome, and Eval metrics through one-question-at-a-time progression before any code is written. This ensures measurable outcomes are defined upfront.

What is the best way to define vision and outcome metrics for a new feature without a PRD?

Defining vision and outcome metrics without a PRD involves applying a prd-lite session workflow that asks only the minimum questions needed to establish what you are trying to achieve and how you will judge success, drafting a preview for approval before final writeout.

Can I resume a goal clarification session if I need to gather more information?

Yes, you can resume a goal clarification session using session-aware NEW, RESUME, and REVIEW handling, allowing you to continue from a saved session or refine an existing goal document without losing your previous progress.

Does the goal elicitation process support web escalation for checking uncertain specs?

The goal elicitation process supports optional web escalation rules that request web checks specifically when needed to fill in uncertain, time-sensitive, or spec-driven parts of the goal during the clarification workflow.

How do I generate and save goal documentation to a specific output path?

To generate and save goal documentation, the workflow drafts a preview and waits for your explicit approval at a writeout gate before writing the finalized vision, outcome, and acceptance criteria to docs/output/goals/<slug>.md.

When do I need to define acceptance criteria for non-trivial repo tasks?

You need to define acceptance criteria for non-trivial repo tasks whenever starting work without existing goal documentation or a PRD-equivalent artifact, ensuring the vision, outcome metrics, and eval layers are established before implementation begins.