What does this skill do, and when should you use it?
seo-profound is the Profound extension skill inside the claude-seo repository. It focuses on tracking how often a brand is cited or mentioned by large language models over time. Its commands expose current citation rates, 30-day trends, surfacing prompts, co-cited competitors, and alerts against a 7-day baseline. It requires a Profound API key and follows the repository’s Claude Code extension workflow.
It queries Profound for a brand’s citation data across ChatGPT, Perplexity, and other supported LLM surfaces. It reports per-LLM current citation rates and 30-day trends, identifies prompts that do or do not surface the brand, lists brands cited alongside it for the same prompts, and returns spike or drop alerts against a 7-day baseline. Metrics should be attributed as “Profound (live, confidence 0.90)” or with the applicable confidence value.
- An SEO team wants ongoing visibility into brand citations on ChatGPT and Perplexity instead of one-off prompt sampling.
- A brand manager needs week-over-week or month-over-month movement in LLM citation rates.
- A content team wants to discover which user prompts surface the brand and which important prompts still miss it.
- A competitive-intelligence team wants to see which brands appear alongside the target brand for the same prompts.
- An operator wants alerts when citation performance rises or falls materially versus the recent 7-day baseline.
How do you install this skill?
- Before use, confirm that PROFOUND_API_KEY is sent only to Profound endpoints and verify Profound reachability from mainland-China networks.
- The skill does not define handling for API failures, quota exhaustion, stale data, or low-confidence results; treat outputs as research signals requiring review.
- The “live, confidence 0.90” convention does not replace independent validation of Profound data sources, sampling methods, or time windows.
- Shell / CLI
- Network access
- Local filesystem
- MCP Server
Profound API keyClaude Code settings file
Install Claude SEO using the repository’s normal installation process, then run extensions/profound/install.sh or install.ps1. Obtain a Profound API key and configure PROFOUND_API_KEY in Claude Code’s ~/.claude/settings.json as required by the skill. The source does not document a separate standalone installation path or additional folder placement instructions.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/AgriciDaniel/claude-seo.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/extensions/profound/skills/seo-profound" ~/.claude/skills/
rm -rf "$tmp"Generated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
In Claude Code, run /seo profound citations <brand> for current citation rates and the 30-day trend; /seo profound prompts <brand> for surfacing and non-surfacing prompts; /seo profound competitors <brand> for co-cited brands; and /seo profound alerts <brand> for changes against the 7-day baseline. Before any tool call, verify that env.PROFOUND_API_KEY exists in ~/.claude/settings.json. Defer Gemini, Google AI Overviews, and AI Mode coverage to seo-seranking, with seo-dataforseo as an additional cross-reference for Google AI Overviews when available.
What are this skill's strengths and limitations?
- Provides continuous time-series tracking for week-over-week and month-over-month analysis.
- Combines citation rates with prompts, co-cited competitors, and anomaly alerts.
- Works as a focused complement to seo-seranking for broader AI visibility coverage.
- Requires source and confidence attribution for reported metrics.
- Requires a Profound API key and extension setup.
- Its native coverage emphasis is ChatGPT and Perplexity, not primary analysis of Gemini, Google AI Overviews, or AI Mode.
- It relies on Claude Code slash commands, settings, and extension conventions, so other clients require adaptation.
- The source does not specify Profound pricing, quotas, latency, or detailed failure handling.
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| Profound AI Citation Tracker this page | 42 · Not recommended | ★ 19k | 23d ago | MIT |
| Social Pulse Monitor | 58 · Recommended | ★ 3k | 13d ago | Apache-2.0 |
| Rank Tracker | 52 · Use with care | ★ 3k | 13d ago | Apache-2.0 |
| Off-Site & AI Referral Analyzer | 59 · Recommended | ★ 3k | 13d ago | Apache-2.0 |
| SERP Analysis | 56 · Use with care | ★ 3k | 13d ago | Apache-2.0 |
Compared with seo-seranking’s on-demand prompt sampling, Profound emphasizes continuous polling and time-series trends. Compared with seo-geo’s end-to-end AI search audit, it focuses on LLM brand citation tracking rather than passage citability and platform-specific optimization.
How did FollowSkills review this skill?
The skill explicitly requires PROFOUND_API_KEY and a local settings check, while repository security material describes user-level credential isolation and permissions; however, it does not disclose request scope, retention, user confirmation, revocation, or rollback, so points are deducted.
Routing, prerequisites, and platform delegation are internally coherent and the happy path is understandable; however, the skill implementation, error handling, API availability checks, and abnormal-input feedback are absent, so the static score remains below the 10-point ceiling.
The audience and four use cases are clear: citation rates, prompts, competitors, and alerts, with explicit delegation for other platforms; however, core reliance on Profound and overseas LLM services leaves Chinese support, mainland-China reachability, non-fit boundaries, and trigger precision underdocumented.
The file provides a name, version 2.2.0, compatibility requirement, installation prerequisite, command table, output attribution convention, and cross-skill delegation; it lacks skill-specific examples, troubleshooting, changelog, maintenance ownership, and update path, so it is not full marks.
The four commands align with the stated LLM citation-tracking goal and expose directly useful trend and alert outputs; however, no real output, execution evidence, or comparative marginal-value evidence is supplied, so the static score is capped below 7.
The skill requires each metric to cite Profound and include a confidence value, and repository CI and tests provide limited supporting signals; however, there are no Profound response fixtures, skill-specific tests, or independent corroboration, so evidence remains thin.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →