Fixing Linked Pages without guessing

Guessing looks like deleting random CMS pages until the sitemap "feels smaller." Fixing looks like reading discovery, deciding the fate of each unexpected public URL, repairing inbound links, and re-running discovery until the graph matches intent.

SlaySlop's check "Discovers public pages reachable from the scanned surface." That is the queue you work.

Outcome you want

Every surprising reachable URL is kept on purpose, redirected with a documented target, or removed from public reachability. Quality checks cover the keepers. The handoff list and the live graph stop arguing.

Boundaries

Permissioned changes on properties you own or manage. Read-only scanners do not edit menus for you. Do not break legal archive requirements just to shrink a report. Not a pentest. Not authorization to scrape unrelated sites because they appeared in an outbound link list—outbound destinations are a different conversation than your own public pages.

Step 1: export the surprise list

Compare Linked pages output to the launch checklist. Sort by whether the URL is linked from global chrome (header/footer) versus one-off content.

# Permissioned notes file
printf '%s\n' "extras-from-discovery.txt"

Step 2: decide with owners

url: /old-pricing
decision: redirect → /pricing
owner: growth
inbound_fix: update 3 blog CTAs

Editors own content links. Engineers own template-level nav. Do not assign everything to "web" and hope.

Step 3: implement redirects or removals correctly

Prefer a single redirect hop. Fix internal links so they stop minting chains. If you remove a page, replace nav entries the same day. Soft 404 templates that still return 200 are not removals.

# Permissioned verification of a redirect
curl -sI "https://www.example.com/old-pricing" | sed -n '1,15p'

HTTP/2 301
location: https://www.example.com/pricing

Step 4: re-discover and cover keepers

Re-run Linked pages. Confirm extras shrank. Run Performance audit and Accessibility audit on remaining key pages—browser-based performance checks and automated accessibility checks on rendered pages—so discovery wins become quality wins.

Pitfalls while fixing

Orphaning pages that sales still emails. Creating redirect loops. Unlinking a URL while ads still spend on it. Forgetting locale twins. Treating outbound vendor links as something you can "fix" by editing someone else's site.

Depth notes

I still reach for curl on redirects, then open the destination in a browser to confirm the rendered page is the one we meant. Discovery plus redirect headers plus a human look beats any single tool glance.

More context

Coordination checklist

[ ] discovery extras exported
[ ] owners assigned per URL
[ ] redirects tested with curl -sI
[ ] inbound internal links updated
[ ] ads/email tools checked for old URLs
[ ] Linked pages re-run
[ ] performance + accessibility on keepers

I still reach for curl first on redirects, then the browser on the destination. That two-step prevents celebrating a 301 to a soft 404.

When extras include files (/brand.zip) or panels, open the security sibling checks instead of pretending it is only an IA problem. Reachability made it urgent; the security catalog names the exposure.

Related next step

When the graph is intentional, keep scanning so new editor links do not silently expand scope. SlaySlop pairs Linked pages with route-attached findings for that follow-through.