# We Audited 4 Pages for Ecosystem Framing: The Cursor-vs-Claude-Code Fix That Mattered

> 4 real Search Console rows for /skills, /commands, /tools/config-converter, and /docs, and the exact title/meta/intro fix shipped for each ecosystem-framing gap.

Canonical: https://www.agentickit.co/blog/cursor-claude-code-framing-audit-fixes
Last updated: 2026-10-29

**Short answer:** We pulled our own Search Console rows for `/skills`, `/commands`, `/tools/config-converter`, and `/docs` and found the same root cause under four different symptoms: the page's title tag and intro were written for one ecosystem while the searcher typed the other, or the branded snippet wasn't compelling enough to click. `/skills` is ranking position 66.7 for "cursor skills" while reading like a Claude-Code-only page — the single biggest leak in this pass. `/commands` sits on page 2 for "cursor commands" with the same ecosystem mismatch but a much closer shot. `/tools/config-converter` is missing the exact phrase ("config maker") people actually search. `/docs` ranks position 3.8 for our own brand name and still gets 0% CTR, which is a title-tag problem, not a ranking problem.

This is the third pass of this audit, after [the first four fixes](/blog/search-console-page-audit-fixes) and [the second pass that sharpened the title-tag-first rule](/blog/snippet-rewrite-playbook-second-audit). `/skills`, `/commands`, and `/docs` already got a title-tag and definition-blurb pass in round two — and in this latest pull they're still showing 0% CTR at roughly the same positions. That's the actual finding here: the round-two fixes answered "what is this page about," but not "which editor is this for," and for queries that explicitly name Cursor, that's the gap that matters. `/tools/config-converter` is new to this pass.

## The four rows, as they actually appeared in Search Console

| Page | Query | Impressions | Avg. position | CTR | What's wrong |
| --- | --- | --- | --- | --- | --- |
| `/skills` | cursor skills | 48 | 66.7 | 0% | Page frames everything as Claude Code skills; a Cursor searcher gets no confirmation this applies to them |
| `/commands` | cursor commands / cursor slash commands | ~11 (combined) | 13.7–32.2 | 0% | Page 2, within striking distance, but title/H1 reads Claude-Code-first |
| `/tools/config-converter` | config maker | 4 | 88.3 | 0% | Searchers say "maker," we say "converter" — no synonym anywhere in the snippet |
| `/docs` | agentickit | 4 | 3.8 | 0% | Strong position on a branded query, generic "Docs" title gives no reason to click |

Four rows, four different queries, and the same underlying issue twice: a page that's technically correct for both ecosystems but written, title-tag-first, for only one of them. AgenticKit supports both Cursor and Claude Code — Claude Code is the primary audience, Cursor is a real secondary one — but if the page copy doesn't say "Cursor" back to someone who searched "cursor skills," Google has no signal to rank it for that query, and the person who does land on it bounces because the first paragraph doesn't match what they typed.

## Fix 1: `/skills` — the 66.7-position leak

48 impressions at position 66.7 means Google is indexing the page for "cursor skills" but burying it past page 6, and 0% CTR means the few people who scroled that far anyway didn't click. The page's existing copy almost certainly opens by describing Claude Code's skill system, then mentions Cursor only in passing (if at all) further down. For a searcher typing "cursor skills," that's a mismatch they can feel in the title and first line, even before reading the body.

The fix:
- Rewrite the meta title and description to name both ecosystems explicitly — something like "Skills for Cursor & Claude Code — 61 Copy-Paste SKILL.md Files," not a title that only says one.
- Add a short, visible intro paragraph (not buried in an FAQ) that states plainly: the catalog installs to both `.cursor/skills/` and `.claude/skills/`, and the trigger-based loading behavior works the same way in both.
- If the page has an install command, show the Cursor-flavored one next to the Claude Code one, not as an afterthought three screens down.

The reasoning is the same lesson from [the Cursor rules vs. skills comparison](/blog/rules-vs-skills-vs-commands-vs-hooks): most searchers typing "cursor skills" don't yet know a SKILL.md is portable across editors. If the page doesn't tell them in the first two sentences, position 66.7 won't improve just because the content underneath is accurate.

## Fix 2: `/commands` — closest to the win, needs the least work

This one is the highest-leverage fix in the batch because it's already on page 2. ~11 combined impressions across "cursor commands" and "cursor slash commands," positions ranging 13.7 to 32.2, 0% CTR. Being within three to four spots of page 1 for some of those queries means the content itself is probably fine — the title tag and the first H1 are what's holding it back, because they likely read "Claude Code commands" with Cursor mentioned only as a secondary editor further down.

The fix: rewrite the title and meta description to name Cursor first or equally ("Cursor & Claude Code Slash Commands: 48 Copy-Paste Examples"), and add an explicit "Cursor slash commands" H2 subsection near the top — not just a note that commands also work in Cursor, but a dedicated section with the `.cursor/commands/*.md` path and a worked example, mirroring what [the dedicated Cursor slash commands post](/blog/best-cursor-slash-commands) already does on its own page. The goal isn't a rewrite — it's giving the snippet something concrete to show a Cursor searcher so CTR has a reason to move off 0%.

## Fix 3: `/tools/config-converter` — the "config maker" synonym gap

4 impressions at position 88.3 is deep, and the reason is almost certainly vocabulary, not relevance: the page is correctly about the thing someone searching "config maker" wants, but every title, H1, and meta description on the page says "converter." Search engines do some synonym matching, but at 88.3 the page isn't even competing — it's not surfacing for the query at all in any meaningful sense.

The fix is a single, low-risk change: add "config maker" as an explicit synonym in three places — the H1 (e.g. "Cursor & Claude Code Config Converter (a.k.a. Config Maker)"), the meta description, and one FAQ entry ("Is this a config maker?" → yes, explain what it converts between). This is the same move from [the title-tag-first rule in pass two](/blog/snippet-rewrite-playbook-second-audit): add the exact phrase the searcher typed, don't rely on the model understanding "maker" and "converter" mean the same thing here.

## Fix 4: `/docs` — a 0% CTR branded query is a title problem, not a ranking problem

This is the odd one in the batch. Position 3.8 on "agentickit" is a strong, near-top-of-page-1 ranking for our own brand name, and it's still getting 0% CTR on 4 impressions. When a branded query ranks well and still doesn't get clicked, the ranking isn't the problem — the snippet is. A generic title like "Docs" or "AgenticKit Documentation" gives someone who already knows the brand name no new reason to click; they typed the brand name because they want something specific (install steps, a quickstart, troubleshooting), and the snippet needs to promise that.

The fix: tighten the title tag to lead with a concrete value prop and a number — "AgenticKit Docs — Install the 34-Agent Team in 2 Minutes" — instead of a generic label. This mirrors how [the 34-agent structure post](/blog/how-i-structured-34-ai-agents) frames the product everywhere else on the site: specific numbers in the headline outperform generic category words, even on a branded query where ranking was never the issue.

## The rule this pass confirms

Across three audit passes now, the fix has split cleanly into two categories, and the split tells you where to spend time:

1. **Deep position, low/no impressions (`/skills`, `/tools/config-converter`):** the content is probably fine, but the title/meta/intro doesn't contain the exact words or ecosystem name the searcher typed. Fix the words first.
2. **Decent position, 0% CTR (`/commands`, `/docs`):** the page already ranks acceptably; the snippet just isn't compelling enough to beat whatever else shows up on that results page. Fix the promise in the title, not the content underneath it.

Neither category needed a content rewrite. All four fixes are title tag, meta description, and a short new section or FAQ entry — the same scope of change as the first two audit passes, which is the actual evidence that "fix the snippet before the content" generalizes past one page.

## FAQ

**Why does a page ranking position 3.8 still get 0% CTR?** Ranking position measures where Google places the result; CTR measures whether the snippet convinces someone to click once they see it. A branded query like "agentickit" ranking well just means Google trusts the page is relevant — it says nothing about whether the title tag gives a searcher who already knows the brand a reason to click instead of scrolling past.

**If AgenticKit installs to both Cursor and Claude Code, why were these pages written Claude-Code-first?** Claude Code is the primary audience for AgenticKit and most of the catalog content was drafted against that framing first, with Cursor support added as a secondary path. That's a reasonable build order, but it leaves exactly this kind of SEO gap — pages that are functionally correct for Cursor searchers but don't say so in the title or first paragraph, so they never surface for Cursor-specific queries.

**Is adding "config maker" to a page as a synonym keyword stuffing?** No — it's naming the thing searchers actually call it, once, in the H1, meta description, and an FAQ entry that answers the literal question. Keyword stuffing is unnatural repetition; this is matching vocabulary a real audience uses, the same fix we made with "cursor commands" vs. "slash commands" in the first audit pass.

**How do you decide whether to fix the title or rewrite the content?** Check impressions and position together. Deep position with few impressions means the content likely isn't matching the query's vocabulary — fix the words in the title/intro. Decent position with 0% CTR means the content is already surfacing — fix the promise in the snippet, not the body.

A fourth pull a day later swapped `/docs` out of the row set and surfaced a new near-miss instead — see [the pass-four audit and the "docs"-query framing fix on `/compare/cursor-rules-vs-skills`](/blog/search-console-audit-pass-four).

## Sources

- AgenticKit's own Google Search Console export, pulled directly for `/skills`, `/commands`, `/tools/config-converter`, and `/docs`, 2026-10-29 (first-party, not published externally)
- [We Audited Our Own Search Console Data: 4 Page Fixes That Actually Moved Position](/blog/search-console-page-audit-fixes) — the first audit pass this one extends
- [The Title-Tag-First Rule: A Second Search Console Audit Pass](/blog/snippet-rewrite-playbook-second-audit) — where the "fix the title before the content" rule was first confirmed
- [Cursor Rules vs Skills vs Commands vs Hooks](/blog/rules-vs-skills-vs-commands-vs-hooks) — the ecosystem-framing distinction behind the `/skills` and `/commands` fixes

_Written by Rakshit Yadav. Published 2026-10-29, last updated 2026-10-29._


## Related
- [config-converter](https://www.agentickit.co/tools/config-converter)
- [rules-grader](https://www.agentickit.co/tools/rules-grader)
- [seo-specialist](https://www.agentickit.co/agents/seo-specialist)
- [on-page-seo](https://www.agentickit.co/skills/on-page-seo)
- [project-conventions](https://www.agentickit.co/skills/project-conventions)
- [/seo-audit](https://www.agentickit.co/commands/seo-audit)
