agentskills.codes
DO

dogfooding

by litago75Source

Critically dogfood the running app by playing real user flows and reporting severity-ranked UX and fun feedback with concrete code-linked fixes. Use when asked to test the app experience, evaluate engagement, or provide product-quality UX critique.

Install

mkdir -p .claude/skills/dogfooding-litago75 && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/15078" && unzip -o skill.zip -d .claude/skills/dogfooding-litago75 && rm skill.zip

Installs to .claude/skills/dogfooding-litago75

Activation

This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.

Critically dogfood the running app by playing real user flows and reporting severity-ranked UX and fun feedback with concrete code-linked fixes. Use when asked to test the app experience, evaluate engagement, or provide product-quality UX critique.
248 chars✓ has a “when” trigger

About this skill

Dogfooding Skill

Hands-on product review workflow for this workspace. Prefer direct interaction over code-only analysis.

When To Use

  • User asks to dogfood, playtest, or critically review UX.
  • User asks if the app is fun, engaging, sticky, or clear.
  • User wants feedback tied to actual interaction evidence.

Inputs

  • Optional focus area: fun, onboarding, retention, accessibility.
  • If no focus is provided, default to fun + usability.

Procedure

  1. Ensure app is running locally.
  2. Play at least one full primary loop.
    • Start screen -> in-game interactions -> win state or terminal state -> replay/reset path.
  3. Stress key interactions.
    • Fast-path completion (speed-run behavior).
    • Error/edge behavior (toggle, undo, reset, back navigation).
    • Post-win continuation behavior and replay quality.
  4. Capture concrete evidence.
    • Cite exact UI behavior observed in browser interactions.
    • Map major findings to code locations when possible.
  5. Report findings in severity order.
    • Focus on bugs, UX friction, fun blockers, retention risks.
    • Keep summaries brief after findings.
  6. Provide actionable fixes.
    • Propose high-leverage changes first.
    • Include smallest effective next steps.

Report Format

  1. Dogfooder verdict (short score + rationale).
  2. Findings ordered by severity.
    • Severity level.
    • Observed behavior.
    • User impact (fun/usability).
    • Relevant file links when available.
  3. What is working well.
  4. Highest-leverage improvements (top 3-5).

Workspace Anchors

Quality Bar

  • Do not provide feedback without interaction evidence.
  • Do not stop at generic advice; tie feedback to observed behaviors.
  • Prioritize issues that reduce fun, clarity, or replay value.
  • Explicitly call out if a critical loop is too easy to game or lacks reward.

Search skills

Search the agent skills registry