Image Quality Enhancer Pro

name: image-enhancer-pro-en slug: image-enhancer-pro-en displayName: Image Quality Enhancer Pro version: "1.0.0" description: "Image enhancement; Pro: Super-res + denoise + batch" author: name: "smart" version: "1.0.0" category: "design" tags: - "图像" - "增强" - "画质"

What you get: a structured workflow that turns a stuck task into a repeatable one, delivering ready-to-use outputs and templates.

Overview

[TRACE 5.0 positioning] This skill is a hands-on tool. It delivers reusable artifacts and templates, not vague advice. The original GitHub skill (ComposioHQ/Image Enhancer) enhances a single image. The Pro version turns it into a reusable quality pipeline: super-resolution upscaling, tiered denoise and sharpening, and batch orchestration, all driven by a diagnose-first method — identify whether the real problem is low resolution, sensor noise, compression artifacts, or exposure drift before choosing the processing order, instead of stacking filters until the image looks artificial.

What I can do

• Diagnose quality: assess resolution, noise type, compression blocking, dynamic range, and color cast, then output a prioritized problem list.
• Design the processing chain: an ordered pipeline (denoise, upscale, sharpen, color correct, export) with the purpose and recommended parameter range for each step.
• Plan upscaling: recommend target size and scale factor by use case (web, print, social, e-commerce hero image), including practical upper limits and distortion risks.
• Orchestrate batches: unified configuration and naming rules for an image set, plus an executable batch script skeleton (ImageMagick or Python).
• Export presets: format, color space, compression quality, and dimension combinations per platform to avoid a second lossy pass.
• Before/after acceptance checklist: detail retention, halo artifacts at edges, natural skin tones, text legibility.

What I cannot do

小葱技能站7w4.net发现了升级插件。

• No inventing detail that never existed; severely blurred or very low-resolution images can only be made more pleasant to look at, not truly restored.
• No use for fabricating evidence, altering identity or financial documents, or misleading edits of a person's likeness.
• No removing watermarks or copyright marks from other people's work.
• No judgment on whether a specific commercial use is licensed; you must confirm rights yourself.

Input specification

Provide: 1. The image (or a description of its dimensions, format, and shooting conditions); 2. Intended use (web, print, e-commerce, social, archival); 3. Optional: target size and resolution, available tools (desktop software, CLI, Python), batch volume, aspect-ratio constraints, and style preference (fidelity-first or appearance-first).

Output specification

Delivered in Markdown: - Quality diagnosis report — issue | severity | likely cause - Recommended processing chain — step | operation | purpose | parameter guidance | cautions - Target spec table — use case | size | format | color space | compression - Batch script skeleton (shell or Python code block) - Naming and directory convention - Acceptance checklist - Risk notes — which steps distort the image when overused

Usage examples

Example 1

Input: An 800×600 old photo with visible noise, to be printed at A4. Output: Diagnosis establishes that denoise must precede upscaling; the chain specifies a denoise strength range, 4x super-resolution, mild sharpening, and print color-space conversion; the spec table gives the pixel dimensions required for A4 at 300dpi; the checklist flags checking faces for a plastic look.

Example 2

Input: 200 e-commerce product photos of mixed sizes, to be standardized as squares with improved clarity. Output: A unified configuration (center-crop and padding rules, common edge length, shared sharpening parameters), a batch script skeleton, an SKU-based naming convention, and a sampling checklist (inspect one in every twenty).

Example 3

Input: A screenshot heavily compressed after repeated re-sharing; text must be legible. Output: Diagnosis identifies blocking artifacts rather than resolution as the main issue; the chain prioritizes deblocking and local contrast, explains the trade-off of upscaling on text edges, and notes that obtaining the original screenshot beats any post-processing.

Error handling

• Extremely low source resolution: state the realistic ceiling honestly and suggest re-acquiring the original.
• Repeatedly compressed source: handle blocking first so artifacts are not magnified by upscaling.
• Unclear use case: default to web specs and attach print parameters as an alternative.
• Inconsistent batch: group by shooting conditions, background, or subject and configure each group separately rather than forcing one preset.
• No tooling available: provide tool-agnostic parameter guidance that can be applied manually in any editor.

Limitations and disclaimer

Image enhancement is estimation and reconstruction; results may differ from the real scene and are unsuitable for forensic, medical, or security identification purposes where fidelity is required. Ensure you hold the rights to the images you process and obtain consent where a person's likeness is involved. No warranty is given for any specific software's output; parameter suggestions must be tuned to your actual material.

Pro upgrade roadmap (Pay Skill reserved)

Billing will be enabled once the individual-creator payment channel launches. Free tier: single-image diagnosis, processing chain, one export spec, basic batch script skeleton. Pro tier: multi-platform export preset library (e-commerce, social, print, video thumbnails), large-batch orchestration with resumable job templates, parameter recipe library by material type (portrait, landscape, product, document scan), automatic before/after comparison reports, color management and print proofing checklists, and priority support. Upgrade path: one-click switch to Pro as soon as billing opens.

Standard 5-step workflow

1. Prepare input: collect requirement keywords and business context.
2. Configure parameters: set required fields (input, output format, constraints).
3. Run: trigger the skill and follow the documented order.
4. Verify results: check against expected results and boundary conditions.
5. Reuse: save successful templates as shortcuts for next time.

Scenarios and triggers

Use this skill whenever you need, want, or are stuck on: High-frequency scenarios: - Restoring old photos for printing - Standardizing and sharpening e-commerce hero images - Cleaning up noisy phone night shots for publishing - Making text readable in a repeatedly compressed screenshot - Upscaling low-resolution assets for a design layout - Pre-processing images so social platforms degrade them less - Confirming 300dpi pixel requirements before print - Unifying an entire album's size and look - Converting product shots to square format with clean edges - Upscaling a low-resolution frame into a video thumbnail - Improving legibility of faded document scans - Producing multi-platform, multi-size variants of one image - Giving a team a reusable batch processing standard - Deciding whether to denoise or upscale first - Diagnosing which step made the result look artificial - Writing a clear delivery spec for an outsourced vendor - Standardizing naming and folder structure for archives Typical users: designers, e-commerce operators, content creators, photography enthusiasts, and marketing teams handling large asset volumes.

这个 Skill 的文档写得比较专业详细，涵盖了图像增强的各种场景和流程，但可惜只是个「说明书」，没有实际可用的代码或工具。对于想直接用的人来说可能会失望——看了一堆功能介绍，却发现没有具体实现。质量诊断和处理链设计思路不错，但缺少真实示例和可运行脚本支撑，实用性打折扣。