dogfood

Automate consumer-use validation of Aether features with fresh agents.

1|Updated Apr 13, 2026
One-click install
npx skills add https://github.com/iamacoffeepot/aether --skill dogfood-iamacoffeepot
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: dogfood
Source: https://github.com/iamacoffeepot/aether/tree/main/.agents/skills/dogfood
Command: npx skills add https://github.com/iamacoffeepot/aether --skill dogfood-iamacoffeepot

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill automates the validation of Aether features by simulating consumer usage, logging any issues, and optionally providing judgment on rendered output.

Core Features & Use Cases

  • Automated Dogfood Validation: Run the consumer-viewpoint validation on new features or before PRs are undrafted.
  • Issue Detection: Identify potential issues like awkward compositions, missing primitives, or documentation gaps.
  • Task Approval and Execution: Author a task, get approval if needed, and then execute it using a fresh Codex subagent.
  • Judgment on Output: Optionally judge the rendered output with a separate subagent for additional validation.
  • Rollup and Reporting: Summarize the findings, including totals, grouped friction, artifact verdicts, and soft holds, and persist the evidence.

Quick Start

Run the dogfood validation for the feature 'Feature XYZ' with the PR number 12345 using the command: dogfood validate -i 12345 -p 'Feature XYZ'.

Frequently Asked Questions about dogfood

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

FAQPage Schema
How do I automate consumer-use validation for Aether features before a PR is undrafted?

You can automate consumer-use validation by running a workflow-driven test with fresh agents that consume public surfaces, log friction, and optionally judge rendered output to report category-specific issues and artifact verdicts.

What is feature validation using fresh agents and how does it detect issues?

Feature validation using fresh agents simulates consumer usage on public surfaces to detect potential issues like awkward compositions, missing primitives, or documentation gaps, summarizing findings into grouped friction and artifact verdicts.

What inputs do I need to run an automated dogfood test on a feature scope?

Running an automated dogfood test requires a feature issue or scope, public surface pointers, and an optional task object to perform the workflow-driven test and report findings.

Can I get a separate judgment on rendered output during feature validation?

You can optionally judge rendered output with a separate subagent during feature validation, providing an additional layer of validation through artifact verdicts in the final rollup report.

How do I run a dogfood validation for a specific PR number and feature name?

Run the dogfood validation using the command `dogfood validate -i [PR number] -p '[Feature Name]'` to execute the consumer-viewpoint test and persist the evidence.

What does the rollup report include after an automated consumer-viewpoint test?

The rollup report includes totals, grouped friction, artifact verdicts, and soft holds, summarizing all findings and persisting the evidence from the consumer-use validation workflow.