本文へ移動
cccskills
無料GitHub で公開

measure-chat-flicker

Measure whether the chat jumps while a turn runs by sampling the Working line's position over CDP and counting upward moves. Use when changing step rows, live status lines, folding, or anything that can shrink content during a run.

インストール方法を見る

含まれるファイル(2)

  • SKILL.md1.4 KB
  • scripts/flicker.mjs1.8 KB

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

Measure chat flicker

Content growing below is normal. Content shrinking during a run (a live line removed, a group closing) moves everything up, and that reads as flicker. This counts those upward moves.

export CDP_URL=http://127.0.0.1:<cdp port>
M=.opencode/skills/measure-chat-flicker/scripts/flicker.mjs
node .opencode/skills/drive-desktop-cdp/scripts/cdp.mjs new-session
node $M arm
node .opencode/skills/reproduce-chat-states/scripts/repro.mjs sequential-commands
node .opencode/skills/drive-desktop-cdp/scripts/cdp.mjs wait-idle
node $M report      # { upwardJumps, downwardMoves, samples }

Note: repro.mjs opens its own new session, so arm after it if you need the arm to survive the navigation, or run send directly.

Reading the result

  • upwardJumps during the run should be 0. One at the very start (the first step replacing "Starting…") is acceptable.
  • Reference: before the live step moved onto the group's own line, five sequential commands gave 6 upward jumps of about 22px in 27s; after, 1.
  • Run the same prompt before and after a change and compare. Keep the window size fixed between runs.

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

同じリポジトリのスキル

概要と使いどころ

Add a new feature, put work behind a feature flag, roll something out (per organization, for everyone, cloud or self-hosted), revert or kill a feature, or make a big change to existing behavior such as a new engine. Use BEFORE writing the feature code, and when finishing or removing a rollout.

日本語の概要は準備中です。原文の説明を表示しています。

different-ai/openwork2.4万2026年10月11日 更新

Local OpenWork Electron browser automation with CDP. Use when driving a local Electron dev app, browser_list, browser_snapshot, browser_eval, composer automation, or local UI smoke tests.

日本語の概要は準備中です。原文の説明を表示しています。

different-ai/openwork2.4万2026年10月11日 更新

Flag customer, prospect, partner, or outside-person identities in the diff. Reported in the Warden security summary.

日本語の概要は準備中です。原文の説明を表示しています。

different-ai/openwork2.4万2026年10月11日 更新

Connect the OpenWork MCP Gateway (https://api.openworklabs.com/mcp/agent) to Claude Code, Codex, Gemini CLI, Cursor, VS Code, Claude Desktop, or ChatGPT so the agent can use the user's OpenWork organization skills, plugins, and connections.

日本語の概要は準備中です。原文の説明を表示しています。

different-ai/openwork2.4万2026年10月11日 更新

Screen a pull request from an outside contributor for hidden, obfuscated, or supply-chain-risky changes before any of its code runs.

日本語の概要は準備中です。原文の説明を表示しています。

different-ai/openwork2.4万2026年10月11日 更新

Create an OpenCode plugin for OpenWork. Scaffolds the plugin file with the correct API shape, tool definitions, and hook registration. Use when the user asks to 'create a plugin', 'write a plugin', or 'make a plugin that does X'.

日本語の概要は準備中です。原文の説明を表示しています。

different-ai/openwork2.4万2026年10月11日 更新

different-ai のスキルをすべて見る

このスキルの問題を報告する