Перейти к содержимому
TakoAPI
Назад к навыкам

vajra

от minhyeong112

Web & Frontend Development
345Число загрузок по данным clawskills.sh
2Звёзды
0Просмотры
0 лайковВойдите, чтобы поставить лайк навыку
—Ещё нет оценок
Ваша оценка

Кратко

Analyze URLs, YouTube videos, tweets, or text for quality, bias, and reliability using the Vajra API (vajra.to).

Описание

Что делает этот навык

Vajra analyzes URLs, YouTube videos, X/Twitter posts, and raw text for epistemic quality, bias, and reliability via the Vajra API. It returns a 1-10 quality score, bias assessment, key takeaways, flagged warnings, and a public permalink to the full report. Results are cached server-side, so repeated submissions of the same URL cost zero credits.

Returns a scored, structured verdict with bias detection and claim warnings in a single API call, replacing manual cross-referencing across multiple fact-checking sites.

Когда его использовать

  • Checking a news article for bias before sharing it
  • Scoring a YouTube video's reliability on a contested topic
  • Evaluating whether a viral tweet contains questionable claims
  • Assessing the epistemic quality of a long-form blog post
  • Getting a structured verdict on a primary source before citing it

Пример сценария

ВВОД

User asks: Can you analyze this article for bias and reliability? https://example.com/news-article

АГЕНТ
  1. 1

    Sends POST request to Vajra API with the URL and type set to 'url'

  2. 2

    Waits 10-60 seconds for analysis to complete

  3. 3

    Parses the response for quality_score, bias_level, tldr, verdict, key_takeaways, and warnings

  4. 4

    Formats and presents the results with the public permalink

ВЫВОД

Article scores 6/10 for quality, moderate bias detected, three key takeaways extracted, one warning flagged, full report at vajra.to/a/UUID

Требования

  • VAJRA_API_KEY environment variable
  • Vajra account at vajra.to/dashboard (free tier available)

Установка

clawhub install minhyeong112/vajra

Выполните команду CLI или вставьте промпт в Claude Code, Cursor, Codex — любой кодинг-агент, которым пользуетесь.

Ссылки