I Built a Political Accountability Agent That Queries Sanity

说明与用途

True Oath 使用 Sanity 构建政治承诺的结构化账本，将其与来源、证据和评估关联起来。

问题

Voters need a reliable way to audit political promises across scattered sources, which keyword search cannot reliably link.

结果

True Oath creates a navigable audit trail for political promises, preserving source links, dates, confidence, and unresolved conflicts, and demonstrates how Sanity can support structured accountability data.

工作流

• Define promise, manifesto, government, source, evidence, milestone, indicator, assessment, claim, integrityEvent schemas in Sanity Studio.
• Populate the production dataset with 15 public sources, 4 promises, 4 evidence records, 4 milestones, 4 outcome indicators, 4 independent assessments, 3 competing claims, 1 manifesto, 1 government record, and 4 integrity events.
• Deploy Next.js web app to Vercel and Sanity Studio to https://true-oath.sanity.studio/ and https://true-oath.vercel.app/.
• Configure a Path One agent to query the True Oath Sanity Context MCP endpoint, filtering on the five accountability document types.
• Use the agent to retrieve promises, follow source references, compare claims, and answer audit questions such as status of 2022 promises, evidence presence, conflicts, and unverifiability.

证据

True Oath is a political accountability ledger starting with Australia and designed to expand to other countries.

来源

The interface presents each promise with a status such as Kept, Partially kept, In progress, Not started, Broken, Reversed, or Unverifiable.

来源

The Sanity dataset contains the first independently collected Australia corpus: 15 public sources, 4 promises, 4 evidence records, 4 milestones, 4 outcome indicators, 4 independent assessments, 3 competing claims, 1 manifesto, 1 government record, and 4 integrity events.

来源

The project currently builds with: cd web && npm run build && npm run lint cd sanity && npm run build

来源