Website Improvement Prd logo

Website Improvement Prd

Community
luongnv89
website-improvement-prd

Generate an approval-gated improvement PRD from a report and baseline, with evidence-backed what/why/value changes and measurable targets. Use for rebuild proposals. Don't use for task breakdown, coding, implementation, or unapproved persistence.

Overview

Publisherluongnv89
Repositoryskills
Skill namewebsite-improvement-prd
Stars
124
Forks
18
Bundled files
Instructions only
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by luongnv89 on GitHub. Read the source before you install it.

Installation

Install the Website Improvement Prd AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/luongnv89/skills.git /tmp/skills
mkdir -p .claude/skills
cp -r /tmp/skills/skills/website-cloner/website-improvement-prd .claude/skills/website-improvement-prd
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Website Improvement Prd in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Website Improvement Prd on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Website Improvement Prd is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Website Improvement PRD

Turns an approved end-user report into a full improvement proposal with measurable value metrics. Writes prd.md after user approval.

When to Use

Trigger when the user asks to:

  • Propose website improvements with expected impact
  • Create a PRD for a site rebuild or improvement
  • Plan improvements with measurable metrics

Do not use for implementation planning or coding — those are separate phases.

Workflow

1. Read the approved report (Phase 2) and analysis (Phase 1)
2. Identify improvement opportunities per dimension
3. For each change: define what, why, and expected value
4. Assemble into prd.md
5. Present to user for review
6. Incorporate edits (loop until approved)
7. Persist prd.md

Output: prd.md Structure

markdown
# Improvement Proposal: <site name>
**Source URL:** <url>
**Date:** <date>
**Version:** 1.0

---

## Executive Summary

2–3 sentences summarizing the main improvement themes and expected impact.

## Current State

Brief summary of the baseline from Phase 1 analysis:
- Category and audience
- Current performance highlights
- Current SEO standing
- Current security posture

## Proposed Improvements

### Change 1: <Title>

**What:** Specific, scoped description of the change.
Example: "Replace the current hero section with a centered layout featuring a clear headline, subtext, and primary CTA button above the fold."

**Why:** Problem being solved, evidence from Phase 1 analysis.
Example: "The current hero burying the CTA below two content sections. Users must scroll past unrelated content before seeing the sign-up option (Phase 1 UI/UX friction point: 'CTA not immediately visible')."

**Expected Value:** Measurable impact.
Example: "Estimated +15–25% improvement in CTA click-through rate. Reduced scroll depth requirement from 600px to 0px for primary action."

---

Repeat for each change. Group changes by dimension:

### UI/UX Improvements
### Performance Improvements
### SEO Improvements
### Security Improvements
### Style Enhancements

## Metrics Summary

| Metric | Current | Target | Delta |
|--------|---------|--------|-------|
| LCP | 4.3s | ≤ 2.0s | -53% |
| CLS | 0.18 | ≤ 0.05 | -72% |
| SEO Score | 62/100 | ≥ 90/100 | +45% |
| Page Weight | 2.1 MB | ≤ 800 KB | -62% |

## Next Steps

Reference Phase 4 (implementation plan) — the approved proposal feeds into `tasks.md` for phased execution.

---

*This proposal is based on analysis of a single crawl of the source site. Actual improvements may vary based on implementation decisions.*

Step 1: Read Inputs

Read the approved report and analysis:

Read file <path-to-report.md>
Read file <path-to-analysis.json>

If either is missing, ask for paths. The orchestrator should have already produced both.

Step 2: Identify Improvements

For each dimension, identify specific, actionable improvements:

  • UI/UX: Layout changes, navigation fixes, CTA improvements, content restructuring
  • Performance: Image optimization, lazy loading, code splitting, font optimization, caching
  • SEO: Meta tag fixes, heading structure, structured data, alt text, canonical URLs
  • Security: HTTPS hardening, security headers, mixed content fixes
  • Style: Typography improvements, color palette refinement, spacing consistency, motion polish

Each improvement must include:

  1. What — specific, scoped change description
  2. Why — problem + evidence from Phase 1
  3. Expected Value — measurable impact statement

Step 3: Compute Metrics Summary

Compare current metrics from Phase 1 against realistic targets:

CurrentTarget Logic
LCP > 3sTarget ≤ 2.0s (good) or ≤ 1.5s (excellent)
CLS > 0.1Target ≤ 0.05 (good)
SEO < 80Target ≥ 90
Page weight > 1 MBTarget ≤ 800 KB

Step 4: Write Draft prd.md

Assemble the proposal using the structure above.

Step 5: Present for Review

Present the draft to the user:

"Here is the improvement proposal. Please:

  1. Approve — save as prd.md
  2. Edit — specify what to change
  3. Regenerate — start over with different focus"

Step 6: Incorporate Edits (loop)

If edits requested:

  • Update the proposal
  • Re-present
  • Repeat until approved

Do not persist until explicit approval.

Step 7: Persist prd.md

Write to output path:

bash
printf '%s\n' "$PRD_CONTENT" > "$OUTPUT_PATH"

Default: $PROJECT_DIR/prd.md or ~/workspace/clones/YYYY_MM_DD_slug/prd.md.

If $ARGUMENTS includes --output <path>, use that.

Confirm:

prd.md saved to: <absolute-path>
STATUS: approved

Return Contract

When invoked by the website-cloner umbrella (Phase 3 gate), the orchestrator gates Phase 4 on this skill's outcome. The contract:

OutcomeSignal
approvedprd.md exists at the resolved output path AND final line of stdout reads STATUS: approved
pendingno prd.md written; final line reads STATUS: pending (user still iterating)
abortedno prd.md written; final line reads STATUS: aborted (user declined)

The orchestrator MUST NOT advance to Phase 4 unless the outcome is approved. A standalone invocation may ignore the status line, but the file-existence rule still holds: no approval, no prd.md.

Acceptance Criteria and Expected Output

Verify the complete proposal before requesting approval:

  • Every proposed change contains a specific What, evidence-backed Why, and measurable Expected Value.
  • Every baseline issue from the approved report or analysis maps to a proposal change or an explicit out-of-scope rationale.
  • Metric targets include units, current values, realistic target logic, and correctly computed deltas; label estimates rather than presenting them as guarantees.
  • prd.md contains executive summary, current state, grouped improvements, metrics summary, next steps, and source-data caveat.
  • The expected result is an approved non-empty markdown file plus exactly one final STATUS: approved; pending or aborted outcomes write no file.

Step Completion Report

text
◆ Website Improvement PRD
··································································
  Inputs validated:     √ pass | × fail ([reason])
  Changes evidenced:   √ pass ([count])
  Targets measurable:  √ pass
  User approved:       √ pass | × pending
  prd.md saved:        √ pass ([absolute path]) | — not approved
  Result:              PASS | BLOCKED | FAIL

Report PASS only when the return contract's file and final status-line conditions both hold.

Edge Cases and Error Handling

FailureBehavior
No input files providedAsk for report.md and analysis.json paths
Invalid input formatReport error and ask for valid files
Report and analysis conflictCite both values and ask which approved baseline governs
Missing baseline metricPropose a qualitative change or mark the target unavailable; never invent a number
User never approvesKeep looping; do not auto-save

Frequently asked questions

What does the Website Improvement Prd AI skill do?

Generate an approval-gated improvement PRD from a report and baseline, with evidence-backed what/why/value changes and measurable targets. Use for rebuild proposals. Don't use for task breakdown, coding, implementation, or unapproved persistence.

Why use Website Improvement Prd on TypingMind?

Because you install it once and use it with any model. Website Improvement Prd is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Website Improvement Prd in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/luongnv89/skills/tree/main/skills/website-cloner/website-improvement-prd. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Website Improvement Prd?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Website Improvement Prd?

As many as you like. As long as a model supports skills, you can use Website Improvement Prd with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Website Improvement Prd AI skill free?

Yes. It is published on GitHub by luongnv89 under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇