Skip to content

Cited fact drift check with approved edits

Keeps a list of facts your pages quote from other sites, such as a fee, a limit or a version number. Each week it rereads the source, compares it to the last value and shows you every file that still has the old number before it changes any of them.

MarketingWorks with Claude Code · CursorCommunity example
Build your promptSee requirements

What it does

  • Opens each source page and pulls the value with an exact quote
  • Compares it to the saved baseline from the last run
  • Checks that your own files still contain the expected value
  • Lists every stale match with the old and new value
  • Applies only the edits you approve

Your approval boundaries

Approval required

These actions always wait for your yes:

  • Editing any file
  • Publishing or committing the change

You decide what happens next.

More about this workflow

This setup is for marketers and solo creators who need steady output without losing their voice. It takes a repeat chore off your plate and hands back something you can check quickly.

First, it opens each source page and pulls the value with an exact quote. Next, it compares it to the saved baseline from the last run. Then it checks that your own files still contain the expected value. After that, it lists every stale match with the old and new value. Finally, it applies only the edits you approve. It works with your web fetch and docs site, each on its own login with limited access.

You get the prep work done and waiting for you. The agent stops and asks for your yes before editing any file and publishing or committing the change, and says plainly if a step cannot be undone.

“It does the fixing, you do the deciding.”
hwajongpark hwajongpark on GitHub · Oct 14, 2026
Before you startWeb fetch · Docs site accessScoped permissionsClaude Code or Cursor
Explore required tools and model recommendations

Tools and access

  • Web fetch

    Handles the web fetch part of this job. Grant public pages, no logged-in sessions.

    API required: No. Built into most agent harnesses. It reads public pages with no login.

    No sign-up needed

  • Docs site

    Handles the docs site part of this job. Grant draft or branch access, not publish.

    API required: Optional. Your docs repo or CMS. Most agents work through a branch or draft.

    No sign-up needed

Official links only. Not sponsored, no affiliate links.

Model recommendations

Best pick

GPT-6.1 Sol

Strong general writer at a mid price.

$2 per million input tokens, $10 per million output tokens, for prompts up to 272K tokens. Longer prompts cost $4 in and $15 out.

OpenAI docs as of October 11, 2026

Cheaper backup

GPT-6 Luna

Very low cost for first drafts and digests.

$0.10 per million input tokens, $0.50 per million output tokens, for prompts up to 272K tokens.

OpenAI docs as of October 11, 2026

Our own picks. Not sponsored, no affiliate links.

Build your prompt

Add your details, then copy the prompt into your tool.

1Your details

Fill these in and the prompt updates as you type.

source url, what to pull, files that cite it

weekly, monthly or quarterly

logins, bot walls

2Your prompt

4 of 4 still empty
Fill these in before you run it: [FACTS TO WATCH: source url, what to pull, files that cite it] [CHECK CADENCE: weekly, monthly or quarterly] [PAGES THAT NEED A MANUAL CHECK: logins, bot walls] [WHO APPROVES EDITS] You are the Cited fact drift check with approved edits agent. Keeps a list of facts your pages quote from other sites, such as a fee, a limit or a version number. Each week it rereads the source, compares it to the last value and shows you every file that still has the old number before it changes any of them. Use only the details filled in above. If any [BRACKETED] item is still blank or unclear, stop and ask for it first. Do not guess. What you do on your own: - Opens each source page and pulls the value with an exact quote - Compares it to the saved baseline from the last run - Checks that your own files still contain the expected value - Lists every stale match with the old and new value - Applies only the edits you approve What you must never do without a yes from a person: - Editing any file - Publishing or committing the change Rules: - Before any step on the list above, stop and ask. Say in plain words what you want to do, who it affects, and whether it can be undone. - If something cannot be undone, say so first, in one short line. - Write drafts, not final actions. A person sends, pays, posts, or deletes. - Keep a short log of every action you took and every source you read. - If you are unsure, ask instead of guessing.
View run instructionsPaste into: Claude Code, Cursor

Your answers stay in this browser only.

Run your workflow

Connect your tools with only the access this setup needs, then follow these steps.

Steps

Do these in order. Test on old data before it touches live work.

  1. 1Set up Web fetch. Keep it to public pages, no logged-in sessions.
  2. 2Set up Docs site. Keep it to draft or branch access, not publish.
  3. 3Paste the prompt where your harness takes instructions. See the note under the Copy button.
  4. 4Do a dry run on old or test data first. Check every draft and every stop point by hand before you let it near live work.
  5. 5Read its log after the first real run. Only widen access once a week of runs looks clean.

Safeguards and trust score

The score describes this example setup’s access, approval boundaries, ability to undo changes, and scope.

Access18 / 30

Can draft things that send or spend, so it sits close to risky actions.

Human approval30 / 30

Waits for a yes on: editing any file, publishing or committing the change.

Undo and logs14 / 20

Proposes changes to live records, so the setup needs a change log and a way back.

Scope18 / 20

Runs on its own login across 2 tools. Fewer connections means a smaller blast radius.

How to raise it

  • Access: Take away its ability to send, post or spend. Let it only write drafts, and have a person press send or pay from their own account +6
  • Undo and logs: Stop it from editing live records. Have it write the change as a proposal you apply, and keep a log of every change with a way to roll it back +3
  • Scope: Run it on one tool only, with its own login that can see just what this job needs +2

Do all of these to reach 91, up from 80.

Want help getting there? Get an AI Trust Audit from Pixeldesigns

Reviewed by Trustcase (example setup). How scoring works

Embed badge

Show this trust score on your site or README. It links back here.

Trustcase trust score badge
Markdown
[![Trustcase trust score for Cited fact drift check with approved edits](https://trustcase.dev/badge/cited-fact-drift-check-with-approved-edits.svg)](https://trustcase.dev/use-cases/cited-fact-drift-check-with-approved-edits)
HTML
<a href="https://trustcase.dev/use-cases/cited-fact-drift-check-with-approved-edits"><img src="https://trustcase.dev/badge/cited-fact-drift-check-with-approved-edits.svg" alt="Trustcase trust score for Cited fact drift check with approved edits" height="20"></a>

More marketing workflows

Other workflows that use the same tools

  • Docs from code

    Engineering. Keeps internal docs in step with the code.

  • Morning AI news brief

    Personal. Gathers the day's AI news with simple scripts, then has the agent pick the best items and write a short PDF brief. It lands in your chat app before 7 AM.

  • Broken link sweep

    Engineering. Checks every link on your docs or site each week and proposes fixes for the dead ones. A person confirms each fix before it goes live.