本文へ移動
cccskills
無料GitHub で公開

scepticagent-prompts

ScepticAgent AI prompt reference. Use whenever the user asks about the prompts, wants to change analysis behaviour, modify what the AI looks for, adjust highlight categories or scoring, understand what instructions are sent to the AI, or tune the output format of analysis or highlights.

インストール方法を見る

含まれるファイル(1)

  • SKILL.md2.4 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。

Prompt locations — all in background.js

buildAnalysisSystem(url, title, content) — lines 658–705

Full page analysis system prompt. Instructs Claude to use web_search to verify claims, check source credibility, find supporting/opposing evidence, identify bias and emotional language. Page content truncated to 10000 chars.

Required output sections (renderer depends on these exact headings):

## Key Claims
## References & Sources
## Supporting Evidence
## Opposing Perspectives
## Bias & Framing
## Emotional Language
## Misinformation Risks
## Overall Assessment

Sources must be cited as inline markdown links e.g. [Reuters](https://reuters.com/...).


buildChatSystem(url, title, content) — lines 707–720

Chat follow-up system prompt. Instructs the agent to answer questions about the page using web search and cite sources as inline markdown links. Page content truncated to 10000 chars.


buildHighlightsPrompt(withSearch) — lines 108–140

Extracts verbatim phrases from the page. Two modes:

  • withSearch: true — includes STEP 1 (web_search for real URLs) / STEP 2 (output JSON). Used in manual "Highlight Page" after analysis is complete.
  • withSearch: false — JSON output only, source_url left empty. Used in the auto fast scan on page open.

Output JSON schema:

{
  "summary": "two sentences — what the article is about, then how emotional/factual/biased it is",
  "category_summaries": {
    "misinformation": "one sentence",
    "emotional": "one sentence",
    "bias": "one sentence",
    "supported": "one sentence",
    "logical": "one sentence"
  },
  "highlights": [
    {
      "phrase": "verbatim text from page (8–180 chars, no quote characters)",
      "category": "misinformation|emotional|bias|supported|logical",
      "note": "max 90 chars explanation",
      "source": "publication name or empty string",
      "source_url": "full URL or empty string"
    }
  ]
}

Constraints: max 25 highlights, phrases must appear verbatim on the page, no markdown code fences in output, no typographic quote characters in phrases.

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

Expert guidelines for Chrome extension development with Manifest V3, covering security, performance, and best practices

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。

Maros112358/scepticagent22026年3月22日 更新

Use this skill when building applications with Gemini models, Gemini API, working with multimodal content (text, images, audio, video), implementing function calling, using structured outputs, or needing current model specifications. Covers SDK usage (google-genai for Python, @google/genai for JavaScript/TypeScript, com.google.genai:google-genai for Java, google.golang.org/genai for Go), model selection, and API capabilities.

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。

Maros112358/scepticagent22026年3月22日 更新

ScepticAgent internal architecture reference. Use whenever the user asks how the extension works, wants to add a new AI provider, add a new highlight category, debug communication between components, understand the agent loop or streaming, or work with provider routing and the Gemini CORS proxy.

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。

Maros112358/scepticagent22026年3月22日 更新

ScepticAgent packaging and Chrome Web Store publishing workflow. Use whenever the user wants to build the zip, package the extension, publish or submit to the Chrome Web Store, update the store listing, or deal with manifest/icon/permission requirements for publishing.

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。

Maros112358/scepticagent22026年3月22日 更新

ScepticAgent UI and styles reference. Use whenever the user wants to change the side panel UI, modify highlight colours, update the markdown renderer, work with CSS architecture, add or change category colours, or understand how the sidepanel renders analysis output.

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。

Maros112358/scepticagent22026年3月22日 更新

Searches multiple web sources, synthesizes findings, and produces cited research reports using delegated subagents. Use when the user asks to research a topic online, search the web, look something up, find current information, compare options, or produce a research report.

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。

Maros112358/scepticagent22026年3月22日 更新

Maros112358 のスキルをすべて見る

このスキルの問題を報告する