SEO indexability snapshot
One HTTP GET snapshots a page's on-page SEO signals — title, meta description, canonical, Open Graph — asserting the tags a crawler actually sees, read-only.
A real, replayable Reelier skill against a live public API. Every replay proves the same thing: A public page serves the core on-page SEO signals in its server-rendered HTML — title, meta description, canonical, and Open Graph — and writes a receipt saying exactly which assertions held.
What this checks
One step: an HTTP GET to a public page, then six assertions against the raw HTML — the server answered 200, and the head carries a <title>, a name="description" meta, a rel="canonical" link, and og:title / og:description Open Graph tags. The page title and canonical URL are bound onto the receipt.
It asserts the HTML exactly as the server delivered it — the same bytes a crawler sees before running any JavaScript — so a page that only injects these tags client-side after hydration will correctly fail, because a crawler wouldn't see them either. The asserts pin the presence of each tag marker, never a value that would rot when you re-title the page.
Not a rank checker. There is no free, read-only, unauthenticated SERP/position API, so this skill never claims to know where a page ranks — it checks the on-page signals that decide how the page is indexed and previewed, never its position.
| Steps | Assertions | Effect | Seed run (2026-07-21) |
|---|---|---|---|
| 1 | 6 | read | ✓ 1/1 passed · 239ms · 0 LLM tokens |
Endpoint it hits:
https://www.reelier.com/skills
The skill file
This is the whole artifact — a markdown file with an assertion on every step. Copy it, check it into your repo, replay it anywhere.
---
name: seo-indexability-snapshot
description: On-page SEO/indexability signals for a public page — title, meta description, canonical, Open Graph — read-only
---
<!-- synced from seldonframe/reelier examples/portfolio — edit there -->
# SEO indexability snapshot
Inputs: (none — this file replays green as-is; see the personalization note)
<!--
Personalization: swap the literal page URL `https://www.reelier.com/skills`
in the action for your own page (the portfolio README's "Point these at
YOUR project" section has the one-command version). The skill grammar has
no default-value syntax for {{var}} holes — an unbound {{var}} is an
explicit error, never a guessed fallback — so this file ships with a
literal that replays green with zero flags.
Honesty note — this is NOT a rank checker. There is no free, read-only,
unauthenticated SERP/position API, so this skill never claims to know where
a page ranks. It checks on-page INDEXABILITY SIGNALS: the head tags that
decide how a crawler indexes the page and how it previews when shared. The
`body contains` asserts test the HTML exactly as the server delivered it —
the same bytes a crawler sees before running any JavaScript — so a page
that only injects these tags client-side after hydration will (correctly)
fail, because a crawler wouldn't see them either.
Endpoint verified live on 2026-07-21: 200; head carries <title>,
name="description", rel="canonical", og:title, og:description.
-->
The assertions pin the PRESENCE of each signal (the tag marker is in the
served HTML), never its exact value — the title text and canonical URL are
bound onto the receipt instead, so re-titling the page or moving its
canonical changes the recorded values without ever breaking the replay.
## Steps
### Step 1 — Fetch the page and check its indexability head tags
- intent: GET a public page and confirm it serves the core on-page SEO signals — title, meta description, canonical, and Open Graph
- action: http.get {"url": "https://www.reelier.com/skills"}
- assert: status == 200
- assert: body contains "<title>"
- assert: body contains "name="description""
- assert: body contains "rel="canonical""
- assert: body contains "og:title"
- assert: body contains "og:description"
- bind: title = body match /<title>([^<]*)<\/title>/
- bind: canonical = body match /rel="canonical" href="([^"]*)"/
- effect: readRun it yourself
reelier run takes a local file, so fetch the canonical copy first (or clone the repo). The replay is read-only, needs no account and no keys, and writes its receipt to .reelier/runs/.
# straight from the repo's main branch (the canonical copy):
curl -fsSL https://raw.githubusercontent.com/seldonframe/reelier/main/examples/portfolio/seo-indexability-snapshot.skill.md -o seo-indexability-snapshot.skill.md
npx -y reelier@latest run seo-indexability-snapshot.skill.md
# or from a clone:
git clone https://github.com/seldonframe/reelier
cd reelier && npx -y reelier@latest run examples/portfolio/seo-indexability-snapshot.skill.mdThe receipt
The portfolio README records a real seed run of this skill, replayed with the published CLI on 2026-07-21. Verbatim output:
✓ Step 1 — Fetch the page and check its indexability head tags [passed] 239ms
PASSED: 1/1 steps ok, 0 failed, 239ms total
[measured] 0 LLM tokensAnd the proof keeps accumulating without anyone touching it: a GitHub Action replays all five portfolio skills every 6 hours (plus on demand) — the public Actions tab is the standing audit log. A failing endpoint produces an honestly failing receipt there, never a retried-until-green one.
Point it at your project
Swap the one literal page URL `https://www.reelier.com/skills` in the action for your own page URL. The skill grammar has no default-value syntax for {{var}} holes — an unbound {{var}} is an explicit error, never a guessed fallback — so this file ships with a literal that replays green with zero flags.
sed 's#https://www.reelier.com/skills#https://YOUR-SITE.com/your-page#g' seo-indexability-snapshot.skill.md > my-seo-snapshot.skill.md \
&& npx -y reelier@latest run my-seo-snapshot.skill.md