CONTENT CREATION

Fact-check blog PRs in your content repo and comment on the diff

When a pull request adds or edits Markdown posts, an agent fact-checks the changed claims and leaves a PR comment flagging each unsourced or contradicted statement before…

CategoryContent Creation
Enginepaperclip
Difficultyintermediate
Triggerevent
Steps5
Setup~15 min

How it runs

The automated pipeline, trigger to output.

  • TriggerPull request opened or updated in the content repoGitHubGitHub
  • ActionFetch the diff and keep changed Markdown post linesGitHubGitHub
  • ActionSearch Exa for a source per extracted claimExa
  • LogicSeparate sourced claims from unsourced/contradicted ones
  • OutputPost a PR review comment flagging weak claimsGitHubGitHub

What it does

For teams that keep their blog or docs as Markdown in Git, this workflow brings fact-checking into code review. On every pull request that touches post files, an agent reads the added and modified lines, isolates factual claims, verifies them against the web, and posts a review comment so unsourced statements get caught before merge.

When to use it

Use it when your content lives in a GitHub repo and publishing happens on merge. It fits engineering-adjacent content teams who already review prose in PRs and want claim verification to sit alongside lint and build checks.

How it works

  1. 1A pull request is opened or updated in the content repository, firing the trigger.
  2. 2The workflow fetches the PR diff and keeps only added or changed lines in Markdown post files.
  3. 3The agent extracts factual claims from those lines and searches Exa for a supporting source for each.
  4. 4A decision step separates claims that are sourced from those that are unsourced or contradicted.
  5. 5The workflow posts a PR review comment listing each weak claim with suggested sources, so reviewers can request changes before merging.

Set it up

What you configure once, before turning it on.

  1. 1
    Connect GitHubRepos, issues, pull requests, actions.
  2. 2
    Connect ExaNeural search across the web.
  3. 3
    Set each agent's modelWe leave models unset so you pick the tier — fast + cheap, or top-quality.
  4. 4
    Tune it to your dataEdit the prompts, filters, and field mappings so it matches how your team works.
  5. 5
    Test, then turn it onRun once against a sample, confirm the output, then enable the trigger.

Run this workflow in your colony.

14-day trial. No DevOps. No Sales call. Provisioned in under a minute.