community-automations/visual-regression

研究與情報

PublicClaude 子代理

視覺回歸測試器

一次搞砸的部署,可能把主視覺圖擠出螢幕外,或弄壞定價表,而且往往要等客戶發現才有人察覺。這個測試工具會排定時間為你的頁面截圖。

sonnet4 天ApifyPlaywrightGPT-4 VisionSlack
ClaudeClaude
ROI for
README.md

為什麼選擇此子代理

一次搞砸的部署,可能把主視覺圖擠出螢幕外,或弄壞定價表,而且往往要等客戶發現才有人察覺。這個測試工具會排定時間為你的頁面截圖。

它會將每次截圖與已核准的基準進行比對,然後詢問視覺模型這個變更是真的壞掉還是預期中的內容更新。你會收到 Slack 警示,附上差異圖與判定結果,可忽略的變動則會被過濾掉,讓這個頻道維持可信度。

運作方式

    • Read

      Used at step 01 to kick off the pipeline.

    • Write

      Used at step 01 to kick off the pipeline.

    • WebFetch

      Used at step 01 to kick off the pipeline.

    • WebSearch

      Used at step 01 to kick off the pipeline.

輸出範例

json
// Sample output
// (generated when the pipeline finishes)

Compare the new screenshot to baseline and return {changed: bool, severity, region, likely_cause}; ignore dynamic content zones marked in config.

Unlock the rest

The full agent definition, install snippet, and starter task are gated for community members.

Members get the full `.md` agent file, the npm / pnpm install one-liners, a starter prompt that we've tuned against real runs, and the open-source repo when this automation ships there. One email, magic link, done.