docs(domain-skills): add toolify, humanbenchmark, google-trends, google-autocomplete

[sunflowerleey]

Summary

• Adds 4 new domain-skill notes captured while doing SEO keyword research for a cognitive-training product (Schulte table / visual search / MOT games):
  • toolify.ai: correct /search/<query> URL pattern (the ?q= query-string form doesn't work), plus a scope warning — it's an AI-tools-only directory and returns irrelevant results for non-AI niches (tested on "focus training" / "attention training").
  • humanbenchmark.com: the full test catalog lives on the homepage, not a /tests route (that 404s); stable extraction snippet; current 8-test list documented for future competitive-gap checks.
  • google-trends: the q=term1,term2,... compare-URL trick to get relative search-interest on one normalized 0-100 chart (much better than pytrends, which 429s quickly), rendering/timing notes for the SPA, and a caveat that Rising-query panels get noisy for low-volume/ambiguous seed terms.
  • google-autocomplete: the suggestqueries.google.com/complete/search?client=firefox endpoint for fast, unthrottled long-tail keyword discovery (no auth, no rate limit hit across dozens of calls).

Test plan

• Each URL pattern / endpoint was exercised live via browser-harness / http_get during the research session and outputs were verified (screenshots + JSON responses).
• No code changes, docs-only — nothing to unit test.

🤖 Generated with Claude Code

---

Summary by cubic

Adds four domain-skill docs for keyword research: Google Autocomplete, Google Trends, Human Benchmark, and Toolify search. These notes clarify correct endpoints, compare URL patterns, stable selectors, and scope limits to avoid dead ends.

• New Features
  • google-autocomplete/suggest-api.md: documents suggestqueries.google.com/complete/search?client=firefox JSON endpoint and a quick expansion strategy; notes light limits and a polite delay.
  • google-trends/explore.md: shows the q=term1,term2,... compare URL for a shared 0–100 chart; SPA load timing and Rising-panel caveats; recommends UI over pytrends due to 429s.
  • humanbenchmark/scraping.md: confirms all tests are listed on the homepage (no /tests route) and adds a stable selector; records the current 8-test list.
  • toolify/search.md: correct /search/<query> path pattern (query-string form fails) and a scope warning that this directory isn’t useful outside its niche; adds a simple card selector.

^{Written for commit 5410cfa. Summary will update on new commits.}

[Copilot]

Pull request overview

Adds new domain-skill documentation notes under agent-workspace/domain-skills/ to capture practical URL patterns, scraping/extraction snippets, and workflow caveats for SEO/keyword research sources used with browser-harness.

Changes:

• Documented Toolify’s /search/<query> URL pattern and scope limitations for non–AI-tool niches.
• Added Human Benchmark homepage-catalog scraping notes and a current test list for competitive gap checks.
• Added Google Trends compare-URL technique and SPA rendering/timing guidance.
• Added Google Suggest (autocomplete) endpoint and a small Python helper for fast keyword expansion.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 1 comment.

File	Description
agent-workspace/domain-skills/toolify/search.md	Notes Toolify search URL pattern, scope caveat, and a basic DOM extraction approach.
agent-workspace/domain-skills/humanbenchmark/scraping.md	Documents homepage-based test catalog scraping and enumerates current test set.
agent-workspace/domain-skills/google-trends/explore.md	Captures Trends compare-URL trick, SPA timing notes, and Rising-query interpretation caveats.
agent-workspace/domain-skills/google-autocomplete/suggest-api.md	Documents the Suggest endpoint plus a simple Python snippet and expansion workflow.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[cubic-dev-ai]

2 issues found across 4 files

Prompt for AI agents (unresolved issues)

Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.


<file name="agent-workspace/domain-skills/toolify/search.md">

<violation number="1" location="agent-workspace/domain-skills/toolify/search.md:28">
P3: The documented extraction snippet only returns matching `<a>` nodes, not the card titles it claims to extract. Consumers that use this result as a list of titles will receive DOM elements instead of strings; mapping each anchor to its heading text (and optionally its href) would make the example usable.</violation>
</file>

<file name="agent-workspace/domain-skills/google-trends/explore.md">

<violation number="1" location="agent-workspace/domain-skills/google-trends/explore.md:20">
P2: The example URL contains literal spaces in the `q=` parameter (`term one,term two,...`), but the preceding line explicitly states that spaces need `%20` or `+` encoding. This contradiction will confuse readers and many HTTP clients will reject or re-encode the URL unpredictably. Encode the spaces to match your own guidance, e.g. `q=term%20one,term%20two,term%20three,term%20four,term%20five`.</violation>
</file>

<sub>Reply with feedback, questions, or to request a fix.

Fix all with cubic | Re-trigger cubic</sub>

[cubic-dev-ai]

P2: The example URL contains literal spaces in the q= parameter (term one,term two,...), but the preceding line explicitly states that spaces need %20 or + encoding. This contradiction will confuse readers and many HTTP clients will reject or re-encode the URL unpredictably. Encode the spaces to match your own guidance, e.g. q=term%20one,term%20two,term%20three,term%20four,term%20five.

Prompt for AI agents

Check if this issue is valid — if so, understand the root cause and fix it. At agent-workspace/domain-skills/google-trends/explore.md, line 20:

<comment>The example URL contains literal spaces in the `q=` parameter (`term one,term two,...`), but the preceding line explicitly states that spaces need `%20` or `+` encoding. This contradiction will confuse readers and many HTTP clients will reject or re-encode the URL unpredictably. Encode the spaces to match your own guidance, e.g. `q=term%20one,term%20two,term%20three,term%20four,term%20five`.</comment>

<file context>
@@ -0,0 +1,39 @@
+Comma-separate terms in `q=`, no URL-encoding needed for the comma itself (spaces still need `%20` or `+`):
+
+```
+https://trends.google.com/trends/explore?date=today%2012-m&geo=US&q=term one,term two,term three,term four,term five&hl=en
+```
+
</file context>

Suggested change

	https://trends.google.com/trends/explore?date=today%2012-m&geo=US&q=term one,term two,term three,term four,term five&hl=en
	https://trends.google.com/trends/explore?date=today%2012-m&geo=US&q=term%20one,term%20two,term%20three,term%20four,term%20five&hl=en

[cubic-dev-ai]

P3: The documented extraction snippet only returns matching <a> nodes, not the card titles it claims to extract. Consumers that use this result as a list of titles will receive DOM elements instead of strings; mapping each anchor to its heading text (and optionally its href) would make the example usable.

Prompt for AI agents

Check if this issue is valid — if so, understand the root cause and fix it. At agent-workspace/domain-skills/toolify/search.md, line 28:

<comment>The documented extraction snippet only returns matching `<a>` nodes, not the card titles it claims to extract. Consumers that use this result as a list of titles will receive DOM elements instead of strings; mapping each anchor to its heading text (and optionally its href) would make the example usable.</comment>

<file context>
@@ -0,0 +1,29 @@
+Card titles are inside `h2`/`h3`/`h4` wrapped by an `<a>`; a plain `querySelectorAll("a")` scan filtering for a heading child works, e.g.:
+
+```js
+Array.from(document.querySelectorAll("a")).filter(a => a.querySelector("h2,h3,h4"))
+```
</file context>