Skip to content

Add a page per check, link repo pages to them, and submit to IndexNow - #17

Merged
hsnice16 merged 2 commits into
mainfrom
feat/check-pages
Oct 12, 2026
Merged

hsnice16 merged 2 commits into
mainfrom
feat/check-pages

Conversation

@hsnice16

Copy link
Copy Markdown
Owner

Summary

  • /check and /check/<slug>: one page for each of the 16 checks, with pass rates and example repos from the leaderboard.
  • Repo pages link to more of the site (check names, six similar repos) and their titles say what the score measures.
  • Changed pages are submitted to IndexNow after each scheduled rescore.

Motivation

Search Console, 18 Sep to 9 Oct: 32 clicks from about 1,500 impressions. Repo pages rank at positions 5 to 10 almost only for navigational lookups such as "axios github", at about 0% CTR. The queries the site is about (AGENTS.md, Cursor rules, agent-ready repos) had no page to land on beyond /methodology and /skill. 284 of 371 repo pages are "Discovered, currently not indexed".

Not a roadmap item.

Changes

  • app/check/, lib/check-content/, lib/check-stats.ts, lib/utils/check-path.ts. One new query in lib/db.ts, which is now at the 300-line cap.
  • Check names link to their page from repo pages, /methodology and language hubs. The home page and footer link to them. Sitemap and llms.txt list them.
  • ALTERNATIVES_LIMIT 3 → 6.
  • Repo title: owner/name: AI agent readiness score 83.8/100.
  • scripts/indexnow.ts (bun run indexnow, --since, --dry-run) and a step in scheduled-rescore.yml that runs it five minutes after a push, so the deploy is live before a crawler is sent to it.

Testing

  • biome ci ., tsc --noEmit, bun run test (256 pass), next build (16 check pages prerendered).
  • Manual, desktop only: /check, /check/agents-md, /check/size, a repo page, /methodology, /language/rust, /. /check/agents_md and unknown slugs 404.
  • bun run indexnow --dry-run lists 435 URLs; with --since it lists only entries whose lastmod moved.
  • Not done: mobile widths, dark theme, and a real IndexNow submission. The new workflow step has not run.

Screenshots / screencasts

None attached.

Docs + roadmap sync

N/A for the changelog: no release is being cut. AGENTS.md and README.md updated for the new routes, files and script.

Risks / rollback

  • The "How we score it" text on each check page restates the signal's thresholds by hand. It will go stale if a signal changes without it; AGENTS.md "Adding a signal" now says so.
  • Repo titles change for every repo page, so search snippets will shift as Google recrawls.
  • The IndexNow step is continue-on-error; a failed submission does not fail the rescore.
  • Revert the two commits to roll back. No schema or data change.

🤖 Generated with Claude Code

hsnice16 and others added 2 commits October 12, 2026 16:12
The IndexNow key has been served from public/ since 95ee50a, but nothing
ever submitted a URL, so Bing still learned about changes only by crawling.

`bun run indexnow` takes its URLs from app/sitemap.ts. The rescore workflow
runs it with --since after it pushes, and waits first: the push only starts
the deploy, and a crawler sent to a page before the new build serves it
re-reads the old one.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Search Console (18 Sep to 9 Oct) showed repo pages earning impressions
almost only for "<repo> github" lookups, at about 0% CTR, and no page at
all for the queries the site is about: AGENTS.md, Cursor rules, making a
repo ready for an agent. 284 of 371 repo pages were still uncrawled.

/check/<slug> answers those questions per check: what it is, why agents
need it, how it is scored and how to pass, with the pass counts, per-language
rates and example repos the leaderboard already holds. The prose in
lib/check-content restates each signal's thresholds by hand; a test fails
when a signal has no entry, but nothing checks the thresholds agree.

Check names on repo pages, the methodology and language hubs link to those
pages, and the home page lists them, so authority reaches them from the page
that holds nearly all the external links. Similar repos goes from three to
six per repo page.

Repo titles say what the number is. "owner/name — 83.8 / 100" was matching
searches for the score string itself.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@github-actions

Copy link
Copy Markdown
Contributor

Agent Friendly Code — score diff

Overall: 91.0 → 91.0 (±0.0)

Per agent

Agent Base Head Δ
Claude Code 95.3 95.3 ±0.0
Cursor 85.5 85.5 ±0.0
Devin 94.8 94.8 ±0.0
GPT-5 Codex 96.3 96.3 ±0.0
Kimi CLI 96.1 96.1 ±0.0
Gemini CLI 85.5 85.5 ±0.0
Aider 83.1 83.1 ±0.0
OpenHands 85.4 85.4 ±0.0
Pi 96.6 96.6 ±0.0

No signals changed between base and head.


Posted by agent-friendly-action.

@hsnice16
hsnice16 merged commit 342e5c7 into main Oct 12, 2026
2 checks passed
@hsnice16
hsnice16 deleted the feat/check-pages branch October 12, 2026 10:44

This branch was successfully deployed

1 active deployment
Preview — bd48c089 Deployed Oct 12, 2026 by vercel[bot]
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant