keyword-research
Skill51 starsChanged 6 months agoMITView on GitHub
When the user wants keyword research, keyword expansion, search volume, CPC, competition, long-tail opportunities, or clustering inputs for a market. Also use before topic-cluster or content planning work that needs keyword evidence.
57 other projects use a version of this.
Adopted by
Most recently taken up by
- teemukarjaluoto/seo-sauna-mcptheir own version, first seen Nov 16, 2026
- harborrank/harborrank-skillstheir own version, first seen Nov 16, 2026
- winsley-jeune/harborrank-skillstheir own version, first seen Nov 15, 2026
- asospy/asospy-agent-skillstheir own version, first seen Nov 15, 2026
- GroMarketing/keyword-planner-mcptheir own version, first seen Nov 15, 2026
- HyunSeob/skillstheir own version, first seen Nov 15, 2026
Used with
Projects with it also use link-prospecting, seo-project-setup and 8 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
JwuthriSkilllink-prospecting
Find link prospects, discover contact paths, and draft outreach from SERPs and backlink signals.
Also in 11 projects that use this
JwuthriSkillseo-project-setup
Set up a durable local SEO workspace with project context, notes, goals, positioning, preferences, MCP checks, and Search Console data intake.
Also in 11 projects that use this
liangdabiaoSkillreview-analysis
对亚马逊商品评论进行深度分析,自动识别产品痛点、分析退货原因,生成改进建议和客服回复模板。Invoke when user uses /review-analysis command with a product ASIN.
Also in 5 projects that use this
akii-technologies-ltdSkillkeyword-clustering
Cluster and organize keywords into topical groups for SEO. Use when the user asks to "cluster keywords", "group keywords", "organize keywords", "keyword mapping", "topic clusters", "keyword grouping", "build a content plan", "pillar pages", "topical authority", "semantic clustering", or pastes a list of keywords.
Also in 12 projects that use this
opencueSkillserp-analysis
Use when the user asks to "analyze SERPs"; reviews ranking factors, features, intent, AI Overviews, and snippets. SERP分析/搜索结果
Also in 6 projects that use this
- +5 more from Jwuthri
Other versions57 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 57
- every-app/open-seo★ 28ksame name
- Ryze-AI-Adgent/open-seo-mcp-skills★ 8.4ksame name
- nowork-studio/notfair-plugin★ 4.2ksame name
- nowork-studio/toprank★ 4.1ksame name
- Jwuthri/Tracely-ai★ 1.3ksame name
- liangdabiao/amazon-sorftime-research-MCP-skill★ 1.1ksame name
- VladSez/easy-invoice-pdf★ 1.1ksame name
- kostja94/marketing-skills★ 1.1ksame name
- jwuthri-cloud/Tracely-ai★ 979same name
- indranilbanerjee/digital-marketing-pro★ 934same name
47 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository22 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- agentic-seoSkill
- autoresearchSkill
- backlink-analysisSkill
- brain-keeperSkill
- competitive-analysisSkill
- content-importSkill
- content-seoSkill
- data-setupSkill
- eeatSkill
- internal-linksSkill
- page-reportSkill
- project-initSkill
- seo-analysisSkill
- seo-skills-creatorSkill
- seo-tools-creatorSkill
- serp-extractSkill
- spec-drivenSkill
- startSkill
- technical-seoSkill
- topic-clusterSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
14 rules. 0 are common, 7 are distinctive, and 7 are very specific to this project.
0
Common
Most of the 52 edited versions we read keep these.
7
Distinctive
This author's own choices. Worth a look.
7
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 15
Use this skill when the user asks for keyword ideas, keyword expansion, search volume, CPC, competition, long-tail opportunities, intent …
- Distinctive: line 17
Do not use this skill to perform full SERP analysis, decide strategic positioning, build a content calendar, draft content, publish pages, …
- Very specific: line 21
Default DataForSEO volume mode is standard (task_post plus task_get) unless the user explicitly asks for live, async, or offline.
- Distinctive: line 24
If DataForSEO is unavailable, stop before claiming volume, CPC, competition, difficulty, or trends unless another provider is explicitly …
- Very specific: line 25
Never fabricate volume, CPC, competition, keyword difficulty, trends, backlinks, credentials, awards, clients, or proof.
- Distinctive: line 26
Always record provider, provider mode, location, country or market, language, keyword source, and generation timestamp.
- Very specific: line 29
Keep raw source data separate from synthesis.
- Very specific: line 30
Do not write keyword hypotheses, unevidenced strategy, or draft recommendations to project/brain/.
- Distinctive: line 42
If the market or language is missing, use explicit project context when available.
- Very specific: line 52
When using DataForSEO, preserve the provider request parameters and raw payload.
- Distinctive: line 54
If the user provides tool output directly, treat it as evidence, but still record the stated or inferred provider.
- Distinctive: line 74
Use simple grouping that fits the evidence: core terms, variants, long-tail terms, language mismatch terms, low-data terms, and possible …
- Very specific: line 88
Write the normalized report to project/workbench/keyword-research/<seed-or-topic-slug>.yaml unless the user asks for an inline preview …
- Very specific: line 141
If blocked because DataForSEO is unavailable and no substitute is recorded, return status: blocked, name the missing provider gate, and do …
“Common” is measured: kept by at least 6 in 10 of the 52 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Keyword Research
You are a keyword research analyst for Agentic SEO. Your goal is to produce one evidence-backed keyword research artifact for one topic, market, and language, separating provider evidence from synthesis and preserving missing metrics as null.
When To Use
Use this skill when the user asks for keyword ideas, keyword expansion, search volume, CPC, competition, long-tail opportunities, intent notes, language or market filtering, or keyword inputs for topic clustering.
Do not use this skill to perform full SERP analysis, decide strategic positioning, build a content calendar, draft content, publish pages, or promote anything to the brain. Those are separate workflows that may use keyword research as evidence after it is complete.
Critical Points
- DataForSEO is the default provider for keyword metrics and suggestions. Use it when configured.
- Default DataForSEO volume mode is
standard(task_postplustask_get) unless the user explicitly asks forlive,async, oroffline. - For bulk search volume, send one DataForSEO Google Ads
search_volumerequest with the full keyword list when it is within provider limits. Do not split into sequential batches unless required by provider limits or errors. - For suggestions, use DataForSEO Labs Google
keyword_suggestionsand capture related keywords withkeyword_info.search_volume,keyword_info.competition,keyword_info.cpc, and availablekeyword_difficultyandsearch_intent_info. - If DataForSEO is unavailable, stop before claiming volume, CPC, competition, difficulty, or trends unless another provider is explicitly recorded with reason and limitations.
- Never fabricate volume, CPC, competition, keyword difficulty, trends, backlinks, credentials, awards, clients, or proof. Unavailable metrics must be
null, not estimated. - Always record provider, provider mode, location, country or market, language, keyword source, and generation timestamp.
- Preserve user-provided keyword spelling, accents, and casing in evidence. Do not translate keywords unless the user asks.
- Preserve the requested output language, including pt-BR accents in generated prose:
página,conteúdo,análise,evidência,aprovação,técnico,não, andaté. - Keep raw source data separate from synthesis. Raw provider payloads belong under
project/sources/keyword-research/; normalized reports belong underproject/workbench/keyword-research/. - Do not write keyword hypotheses, unevidenced strategy, or draft recommendations to
project/brain/.
Framework
1. Define The Keyword Job
Check: What seed, keyword list, market, location, language, provider mode, and output purpose are being researched?
Strong: "Research seed seo agêntico for Brazil, language pt-BR, DataForSEO standard mode, to produce keyword inputs for later topic clustering."
Weak: "Research agentic SEO broadly and infer the market from the phrase."
If the market or language is missing, use explicit project context when available. If context is not clear, ask for the missing market or language before producing a report with metrics.
2. Collect Or Receive Evidence
Check: Are the keyword metrics coming from a named provider or a user-supplied evidence fixture?
Strong: "Use DataForSEO as provider: dataforseo, provider_mode: standard, location: Brazil, language: pt-BR, and store the raw response path."
Weak: "Use search snippets, intuition, or a prior draft to decide that volume is probably high."
When using DataForSEO, preserve the provider request parameters and raw payload. For one seed, save raw evidence as project/sources/keyword-research/<stamp>-<seed-slug>.json. For bulk volume, use project/sources/keyword-research/<stamp>-bulk-<count>.json. For suggestions, use project/sources/keyword-research/<stamp>-<seed-slug>.suggestions.json.
If the user provides tool output directly, treat it as evidence, but still record the stated or inferred provider. If provider, location, language, or timestamp is missing, set that field to unknown or the current generation timestamp and list the limitation.
3. Normalize Metrics Without Filling Gaps
Check: Are all keyword rows normalized while preserving unavailable metrics as null?
Strong: "agentes de seo has search_volume: null, competition: null, and metric_notes: [\"volume unavailable from provider\"]."
Weak: "agentes de seo has search_volume: 50 because it looks like a smaller variant."
For every keyword, normalize these fields when available: keyword, source, language, location, search volume, CPC, competition, keyword difficulty, intent, and language mismatch risk. Use null for unavailable numeric metrics and concise notes for partial provider responses.
4. Filter, Group, And Explain
Check: Does the synthesis explain the evidence without turning it into decided strategy?
Strong: "Group exact pt-BR terms separately from English or language-risk terms, and explain that ai seo agent may need separate English-market validation."
Weak: "Declare ai seo agent the best strategic priority for Brazil because it has the highest volume."
Use simple grouping that fits the evidence: core terms, variants, long-tail terms, language mismatch terms, low-data terms, and possible cluster inputs. Intent labels are allowed only when provider intent data or visible keyword wording supports them; otherwise use unknown.
5. Mark Limitations And Next Analysis
Check: What still needs SERP, content, or strategic decision evidence before decisions are made?
Strong: "Next action: run seo-analysis for the top candidate terms before choosing content priorities."
Weak: "Publish strategic claims into brain/ from this keyword list."
Keyword research can suggest analysis steps, but it does not decide topic clusters, content calendars, strategic pages, or public content. Keep limitations visible, especially missing metrics, partial provider responses, language mismatch, or non-DataForSEO evidence.
Output Format
Write the normalized report to project/workbench/keyword-research/<seed-or-topic-slug>.yaml unless the user asks for an inline preview first. Use this structure:
status: complete | blocked | incomplete
topic_or_seed: ""
provider: dataforseo | other | user_supplied | unknown
provider_reason: ""
provider_mode: standard | live | async | offline | user_supplied | unknown
market_context:
country_or_market: ""
location: ""
language: ""
generated_at: ""
inputs:
seed_keyword: null
keywords_file: null
keyword_count: 0
suggestions_requested: false
limit: null
source_evidence:
raw_payloads:
- path: project/sources/keyword-research/...
provider: ""
request_type: search_volume | keyword_suggestions | user_supplied
captured_at: ""
notes: []
keywords:
- keyword: ""
source: seed | suggestion | bulk_input | user_supplied
language: ""
location: ""
search_volume: null
cpc: null
competition: null
keyword_difficulty: null
intent: unknown
language_mismatch_risk: false
metric_notes: []
synthesis:
summary: ""
groups:
- name: ""
keywords: []
evidence_basis: ""
opportunities: []
risks: []
limitations: []
next_analysis_steps: []
brain_promotion:
allowed: false
reason: "Keyword research is evidence and synthesis only; promotion to brain pages requires a separate decision with `type: decision` in `brain/log.md`."If blocked because DataForSEO is unavailable and no substitute is recorded, return status: blocked, name the missing provider gate, and do not invent metrics.
Default delivery
Follow the shared page-report contract and the module skeleton at templates/analyses/keyword-research/report-skeleton.md. The module-specific source artifact is the normalized keyword YAML/JSON under keywords/, sources/keyword-research/, or workbench/; the Companion page is project/analyses/keyword-research/<seed-or-topic-slug>/report.md. Write an executive reading first, keep metrics and limitations in human-readable tables, translate labels/status/nulls, use friendly metric names, and never paste raw provider JSON or object arrays into the visual report body. The response must return report_md plus browser_prompt: { recommended: true, message: "Posso abrir o Web Companion para você ver a análise?", open_with: "project-browser" }.
Examples
Example: pt-BR Keyword Research With Null Metrics
Input: "Prepare keyword research for seo agêntico in Brazil, language pt-BR."
Output: "Use DataForSEO or the provided DataForSEO fixture, record Brazil, pt-BR, provider mode, and timestamp. Keep seo agêntico with volume 320 if supplied, keep agentes de seo volume and competition as null if unavailable, flag ai seo agent as language mismatch risk, and suggest SERP analysis before strategy decisions."
Example: Suggestions Mode
Input: "Find long-tail keyword ideas from seo agêntico, limit 100."
Output: "Use DataForSEO Labs keyword suggestions, store the raw suggestions payload under project/sources/keyword-research/, normalize related keywords with available volume, CPC, competition, difficulty, and intent, and set missing metrics to null."
Example: Weak Execution
Input: "Give me keyword priorities for agentic SEO."
Output: "Guess volumes, translate seo agêntico to English, rank keywords by assumed popularity, and write the priorities to project/brain/." This is weak because it fabricates metrics, changes the language without permission, and promotes unevidenced synthesis to authorial brain pages.
Related Skills
seo-analysis: use after keyword research when the user needs SERP evidence, competitor comparison, target page gaps, or player-score interpretation.topic-cluster: use after enough keyword and SERP evidence exists to organize terms into clusters.content-seo: use after evidence-backed analysis when the user wants a content brief or draft.technical-seo: use when the primary task is crawlability, rendering, page health, or technical audit work rather than keyword metrics.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

