Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add MoizIbnYousaf/marketing-cli --skill openseo-keyword-clusteringgit clone --depth 1 https://github.com/MoizIbnYousaf/marketing-cliWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/moizibnyousaf/marketing-cli/openseo-keyword-clustering)<a href="https://agentmods.dev/skills/moizibnyousaf/marketing-cli/openseo-keyword-clustering"><img src="https://agentmods.dev/badge/skills/moizibnyousaf/marketing-cli/openseo-keyword-clustering/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/moizibnyousaf/marketing-cli/openseo-keyword-clustering"><img src="https://agentmods.dev/badge/skills/moizibnyousaf/marketing-cli/openseo-keyword-clustering.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 1 finding, up to medium
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- medium Excessive Agency · line 72 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00116 | $0.01172 |
| Opus 5 | $0.00058 | $0.00586 |
| Sonnet 5 | $0.00023 | $0.00234 |
| Haiku 4.5 | $0.00012 | $0.00117 |
Grade A, and why
openseo-keyword-clustering scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 11d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 87 lines — stays where its author put it; the contents beside it link to each section on GitHub.
OpenSEO Keyword Clustering
Group keywords into page-level clusters and assign each to an existing URL, a new page proposal, or a do-not-target bucket. Output lands in marketing/seo/clusters/*.md (per-cluster briefs) and feeds seo-content / seo-machine with measured page targets.
On Activation
- Catalog + binding:
mktg catalog info openseo --json --fields configuredand.seo/openseo.jsonforprojectId. No binding →openseo-project-setupfirst. No OpenSEO at all → do lexical clustering with Exa context but label every cluster "SERP-unvalidated." - Keyword source: prefer (a)
brand/keyword-plan.mdpopulated sections, (b)list_saved_keywordswith a tag, (c) GSC live data, (d) seed discovery viaresearch_keywords. Under 10 usable terms → simple map, no clustering ceremony.
OpenSEO MCP Tools
get_search_console_performancewithdimensions: ["query","page"]: the cannibalization truth source — one query splitting impressions across multiple URLs means existing cannibalization, not theoretical risk.get_ranked_keywords: domain/page-driven clustering from exact ranking rows + URLs.get_serp_results: SERP-overlap validation for borderline terms (small batches, ≤10).list_saved_keywords/research_keywords: source sets.save_keywords: apply cluster tags ONLY after confirmation.
Workflow
- Assemble the candidate set from the source hierarchy above; dedupe and drop off-strategy terms.
- Cluster by intent and page type: same SERP intent + similar ranking pages belong together; different intent/buyer stage/SERP format splits. Lexical similarity is NOT evidence of same-page fit.
- Borderline terms: small
get_serp_resultsoverlap check. - Assign each cluster: existing URL (if supplied and fitting), new-page proposal, or do-not-target/later.
- Cannibalization check: GSC query+page data first; otherwise flag where two proposed pages share one intent.
- Write per-cluster briefs to
marketing/seo/clusters/<cluster-slug>.md: page type, searcher problem, required sections, internal-link targets, priority. - Optionally tag clusters with
save_keywordsafter confirmation. - Hand off:
seo-contentfor single pages,seo-machinefor programmatic batches — both now have measured targets instead of vibes.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 11d ago First seen · 87 lines · 116 tokens per session scan A 2d5228e397d7
openseo-keyword-clustering is a skill published in the GitHub repository MoizIbnYousaf/marketing-cli (31 stars, last pushed 24d ago), licensed MIT. It adds 116 tokens to every session and 1,172 once invoked, about $0.0006 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
citedy-content-writer
From topic to published blog post in one conversation — generate SEO- and GEO-optimized articles with AI illustrations and voice-over in 55 languages, create social media adaptations for 9 platforms, set up automated content sessions, and manage product knowledge base. End-to-end blog autopilot. Powered by Citedy.
citedy-trend-scout
Find what your audience is searching for right now — scout X/Twitter and Reddit for trending topics, discover and deep-analyze competitors, and find content gaps. Combine social signals with SEO intelligence. Powered by Citedy.
launch-tier-planner
Use when the user asks to "plan my launch tier", "how big should this launch be", or "build a launch risk register with kill criteria"; produces a tier decision (Tier 1 flagship all-channel / Tier 2 targeted / Tier 3 changelog-level), a launch-type declaration (new-product / feature / relaunch / partnership with…
narrative-drift-monitor
Use when the user asks to "check if our surfaces have drifted from the canon", "watch for competitor repositioning", or "define when we should reposition"; produces a drift report — self-drift per flagship surface vs the narrative-registry canon over time (via wayback.py, change history Measured with as-of dates)…
advocacy-program-designer
Use when the user asks to "design an employee advocacy program", "set up founder-led sharing", or "build a share kit for the team"; produces an advocacy program blueprint in two modes — participation-driven opt-in (default) or top-down assigned with its coercion and authenticity risks flagged — with a voluntary opt-in…
engagement-inbox-manager
Use when the user asks to "triage our comments, DMs, and mentions", "draft replies to this thread", "can we repost this fan post", or "set up inbox SLAs and an escalation path"; produces a ranked triage queue with register detection (sincere / ironic / performative / parasocial, sentiment-inversion table included …