How to use the Profound LLM Citation Tracker agent skill

What the seo-profound skill makes the agent do

A skill is a folder of instructions that an AI agent loads when a task matches it. This one focuses on brand visibility inside AI answer engines. Profound continuously polls these engines and publishes time-series data, so the agent can report week-over-week and month-over-month changes rather than a single snapshot.

The skill routes four commands. Each figure the agent reports is labelled with its source, for example "Profound (live, confidence 0.90)".

• /seo profound citations <brand>: current citation rate per LLM plus a 30-day trend
• /seo profound prompts <brand>: top prompts that surface, or fail to surface, the brand
• /seo profound competitors <brand>: brands cited alongside your brand for the same prompts
• /seo profound alerts <brand>: spike and drop alerts compared with a 7-day baseline

What you need before you install

The skill needs three things: the claude-seo framework installed, a Profound API key, and the Profound install script run from the repository. The key is stored as PROFOUND_API_KEY in ~/.claude/settings.json. The skill checks for this key before any tool call, and without it the skill cannot return live data.

The skill's own description says it requires a Profound API key. Check Profound's terms for pricing and access; this guide does not state them.

• claude-seo framework installed
• A Profound API key, stored as PROFOUND_API_KEY in ~/.claude/settings.json
• Run extensions/profound/install.sh (macOS/Linux) or install.ps1 (Windows) from the repository

Install the skill from the ZIP

Download the ZIP for seo-profound from this page and unzip it. You can install it in one of two places. For all your projects, unzip it into ~/.claude/skills/. For a single project, unzip it into that project's .claude/skills/ folder.

If you use the Claude apps instead of the file system, you can upload the skill under Settings → Capabilities → Skills.

• Personal install: ~/.claude/skills/
• Project install: .claude/skills/ inside your project folder
• Claude apps: Settings → Capabilities → Skills, then upload

Run the Profound install script and set your key

The skill's compatibility note says to set PROFOUND_API_KEY by running extensions/profound/install.sh. Run the script from the claude-seo repository, or install.ps1 on Windows, and make sure the key ends up under env.PROFOUND_API_KEY in ~/.claude/settings.json.

Confirm the key is present in that file before you start asking for reports.

How to trigger the skill with natural requests

You can use the slash commands directly, or describe what you want and let the agent pick the matching command. Name the brand clearly, since the commands take the brand as their argument.

Ask for trends when you care about change over time. The skill is built around week-over-week and month-over-month comparisons, so phrasing like "how has our citation rate moved this month" fits its purpose.

• "Run /seo profound citations for Acme Analytics and show the 30-day trend."
• "Which prompts surface Acme Analytics on ChatGPT, and which ones miss it?"
• "Who gets cited alongside Acme Analytics for the same prompts?"
• "Alert me if Acme Analytics citations spike or drop against the last 7 days."

Safety notes: scripts and network access

This skill runs scripts and uses the network. The install step runs repository scripts, and the commands call the Profound service. Review the install scripts before you run them, and keep your API key out of shared files.

The skill also treats fetched pages, search results, API responses and third-party files as untrusted data. If any of that content tries to instruct the agent, such as asking it to run a command or change a setting, the skill says to report it rather than follow it. Keep that in mind when you read results.

• Install scripts run on your machine; read them first
• Commands send requests to the Profound service using your key
• Treat any instructions found inside fetched content as suspicious

Tips for getting useful results

Use the trend commands for decisions and the prompts command for content work. Citation rate tells you how often you appear; the prompts list tells you which questions you miss, which is where content changes are usually made.

For Gemini, Google AI Overviews or AI Mode, the skill says to use the seo-seranking skill instead. For Google AI Overviews citation rates it also suggests cross-checking seo-dataforseo tools when they are available. For a full AI search audit, the skill hands work back to seo-geo, and for prompt-set design and detecting AI-style patterns in cited content it falls back to seo-content.

• Use alerts to catch sudden changes rather than checking manually
• Compare competitors on the same prompts to see who is cited instead of you
• Pair with seo-seranking for triangulated coverage across engines