AI Skills One-Year: Top 10 Global Cumulative Installs

Image Analysis

Image 1 — Facts (readable text, UI steps, chart data and visible phenomena, organized in screen order):

- Top title (verbatim):

"2026全球 AI Skills 累计安装榜 Top10"

- Subtitle under title (verbatim):

"找技能 · 理清需求 · 写代码 · 美化网页"

- Small right label (verbatim):

"Skill 榜 第二期"

- Layout and overall style:

- A single vertical ranking list (1 to 10), each item contains: rank/icon/skill name/short description/source (username/repo)/install count (blue number at right with a downward arrow icon).

- Gold/silver/bronze medal icons at left for ranks 1/2/3.

- Many items show the same portrait avatar at left; some items show a black triangle or black "A" icon.

- Small-footnote statistics and a timestamp at the page bottom, plus a small signature/label at the lower right.

- Rankings and items (as shown in image, preserving original text/number format):

1. Name: find-skills

- Description (verbatim): 按你的需求，搜索并推荐 AI 技能

- Source/label: vercel-labs / skills

- Installs (displayed): ↓ 380万

2. Name: grill-me

- Description (verbatim): 让 AI 提问，帮你把想法变成具体方案

- Source/label: mattpocock / skills

- Installs (displayed): ↓ 130万

3. Name: grill-with-docs

- Description (verbatim): 把开发想法问题清楚，并记录讨论结论

- Source/label: mattpocock / skills

- Installs (displayed): ↓ 110万

4. Name: improve-codebase-architecture

- Description (verbatim): 找出代码里哪些难维护，并给出改进建议

- Source/label: mattpocock / skills

- Installs (displayed): ↓ 110万

5. Name: agent-browser

- Description (verbatim): 让 AI 帮你点网页、填写表单、截取数据

- Source/label: vercel-labs / agent-browser

- Installs (displayed): ↓ 110万

6. Name: tdd

- Description (verbatim): 先写功能测试，再让 AI 写代码实现

- Source/label: mattpocock / skills

- Installs (displayed): ↓ 100万

7. Name: setup-matt-pocock-skills

- Description (verbatim): 配置项目任务管理，方便 AI 按统一规则开发

- Source/label: mattpocock / skills

- Installs (displayed): ↓ 96.97万

8. Name: frontend-design

- Description (verbatim): Anthropic 官方：帮你设计网页布局和配色

- Source/label: anthropics / skills (see “uncertain content” below)

- Installs (displayed): ↓ 96.68万

9. Name: handoff

- Description (verbatim): 记录需求变更过程，方便将 AI 接着做

- Source/label: mattpocock / skills

- Installs (displayed): ↓ 94.91万

10. Name: triage

- Description (verbatim): 整理项目问题和需求，列出可执行的任务

- Source/label: mattpocock / skills

- Installs (displayed): ↓ 91.35万

- Visual and minor elements (visible phenomena):

- The words "AI Skills" at the top have "AI Skills" in blue; other text is black/gray.

- Each install count at right is preceded by a blue downward-arrow icon.

- Ranks 1–3 have small gold/silver/bronze medal icons at left.

- Items 1 and 5 use a black triangle icon; item 8 uses a black background white "A" icon; many other items show the same portrait photo.

- Footer contains a small statistical note and a timestamp (transcribed below).

- Footer statistical note (verbatim as transcribed):

"统计说明：skills.sh 累计安装量（约数），并非使用人数或全球安装总量；截止 2026.10.09."

(Lower-right may show a small signature/label such as "@格格说" which appears slightly blurred.)

Image 1 — Unclear/uncertain content (itemized):

- The source label for item 8 appears as either "anthropics / skills" or "anthropic / skills": the small text is not sufficiently sharp to determine whether the middle letters form "anthropic" or a pluralized form. For strict citation, verify with the original image file or source page.

- The lower-right signature/label (e.g., "@格格说") is small and slightly blurred; exact spelling should be verified on the source page.

Image 1 — Inferences (based on visible cues, not directly readable):

- The list appears ordered by "cumulative installs" since the title includes that phrase and the right column shows install numbers; top-ranked items have larger install numbers.

- The source fields (e.g., "mattpocock / skills") likely follow a GitHub-style "username / repository" format (image shows slash-separated short texts common in repo attributions).

- "Skill 榜 第二期" suggests this is the second edition of the ranking.

- The repeated portrait likely represents the same person’s avatar used for multiple entries, but the image does not identify that person and no identity claims are made here.

Notes (compliance with image-internal instructions):

- Instructional text inside the image was ignored; only readable content and visible phenomena were transcribed.

- Original numbers, names and short descriptions from the image were preserved; unclear source strings were marked as uncertain.

SkillPI Creation Analysis

Purpose and verifiable content:

- This material is a ranked summary showing the "2026 全球 AI Skills 累计安装榜 Top10", listing skill names, short descriptions, source labels (appearing in a username/repo style) and approximate install counts. It is useful for quickly identifying prominent AI Skills and their described use-cases. Original source URL provided by the user: https://www.xiaohongshu.com/explore/6ac8f64b000000000a0202ad?xsec_token=YBTPDMMUX25_VOy5nmmw6XaNsoRdkzbVhAz4t4gUIn2-E%3D

Image evidence key points to verify:

- The image shows each rank, the skill names (e.g., find-skills, grill-me, agent-browser), source labels (e.g., vercel-labs / skills, mattpocock / skills), and install counts in the format shown (e.g., "↓ 380万").

- The footer states data come from skills.sh and are approximate cumulative installs with a timestamp "截止 2026.10.09", indicating a specific cutoff date for the snapshot.

- Some text in the image is not fully legible (notably the source string for item 8 and a small signature), and should be confirmed against the original page or image for exact spelling and links.

Practical steps to use or validate listed Skills (based on the image and summary descriptions):

- find-skills: Describe the task you want to accomplish (e.g., build a webpage, write tests)→ use find-skills to search for matching Skills→ open the listed source/repo to read installation and usage instructions→ install or enable the Skill per its documented steps.

- grill-me: When you have a rough idea, run the Skill to prompt iterative questions from the AI→ answer and refine the scope and constraints→ produce a clarified plan to hand off to development or other Skills.

- grill-with-docs: Use during planning discussions to capture key concepts, decisions, and rationale→ produce referenceable documentation for future review.

- improve-codebase-architecture: Run a scan or analysis (as provided by the Skill) to locate complex or hard-to-maintain areas→ receive a visual report and prioritized suggestions for refactoring.

- agent-browser: Provide the necessary platform permissions (where supported) for automated browser actions (open pages, click buttons, fill forms, capture screenshots, extract page data)→ use results for site checks or data capture, observing platform security and permission constraints.

- tdd: Define expected tests first (test cases)→ have the AI generate code to satisfy those tests→ iterate using tests as verification checkpoints.

Reproduction prerequisites and cautions:

- Access to a platform or environment that supports Agent Skills is required; the summary states that Anthropic announced the concept in Claude on 2025-10-16. To install or run these Skills you typically need to follow the source repo or platform-specific integration steps.

- The image’s source labels resemble GitHub-style "username / repo" notations; actual installation and usage depend on each repository’s README and any platform-specific adapters or permissions.

- The footer clarifies install counts are approximate and time-bound (skills.sh, cutoff 2026.10.09). For up-to-date or exact figures, consult skills.sh and the individual source repositories.

Items requiring verification before citation or replication:

- The source label for item 8 might be "anthropics / skills" or "anthropic / skills"; the screenshot is not sharp enough to confirm. Verify on the original image or source page to cite the correct repository or organization.

- The lower-right signature/handle visible in the image is slightly blurred (e.g., "@格格说"); confirm the exact author/handle on the source page if attribution is needed.

- Install counts shown are labeled as approximate; to rely on numbers for analysis, fetch data directly from skills.sh or from each repository’s official statistics.

Verification and action recommendations:

- To replicate or try a listed Skill, navigate to the source label shown in the image (e.g., vercel-labs / skills, mattpocock / skills, vercel-labs / agent-browser) and follow the repository README and installation instructions.

- For statistical claims, check skills.sh (referenced in the image footer) and the respective repository pages for the latest numbers and timestamps.

Note: This interpretation relies solely on the provided summary and the transcribed image text; no external repository or platform pages were accessed to validate implementation details. All procedural suggestions are limited to the capabilities described in the image and summary and should be validated against the actual Skill repositories and the supporting platform before execution.