All skillsCompetitors
Competitor Gap Analysis
Where rivals win on your prompts
Install this skill
Skills are SKILL.md files your agent loads for specialized workflows. Most tools discover them under .cursor/skills/, .agents/skills/, or a global skills folder. See the Agent Skills spec.
- Create
.cursor/skills/competitor-gap-analysis/in your repo (or user skills folder). - Save the skill file as
.cursor/skills/competitor-gap-analysis/SKILL.md. - Connect AI SEO Tracker MCP from your project Settings -> Agent page (OAuth, no secrets in
mcp.json). - In your agent (Cursor, Claude Code, etc.), invoke
/competitor-gap-analysisor describe the task — the agent applies the skill when relevant.
Quick install (terminal)
mkdir -p .cursor/skills/competitor-gap-analysis/ && curl -fsSL "https://aiseotracker.com/api/skills/competitor-gap-analysis" -o .cursor/skills/competitor-gap-analysis/SKILL.md
SKILL.md
Full skill file your agent reads.
---
name: competitor-gap-analysis
description: Compares your brand vs competitors on shared prompts using snippets and ranks. Use for competitive analysis, battlecards, or "why does X rank above us in ChatGPT".
---
# Competitor Gap Analysis
## Prerequisites
AI SEO Tracker MCP must be connected before running this skill.
1. Open AI SEO Tracker -> Project Settings -> Agent (`/p/{projectSlug}/settings/agent`).
2. Copy the MCP URL shown on that page (`https://aiseotracker.com/api/agent/mcp` for this environment).
3. Add it to your MCP-capable client. Cursor example:
```json
{
"mcpServers": {
"aiseotracker": {
"url": "https://aiseotracker.com/api/agent/mcp"
}
}
}
```
4. Reload/restart your MCP client so it discovers the new server. In Cursor, reload the window after saving `.cursor/mcp.json`.
5. In Cursor, open Settings -> MCP, find `aiseotracker`, and click **Connect**.
6. Approve the AI SEO Tracker OAuth prompt in the browser. No API key or static token is required.
7. If a tool call returns `agent_auth_required`, the client has not completed OAuth yet. Reconnect/reload MCP from the Agent settings page and retry.
Available MCP tools:
- `get_project_sources` - citation URLs and domains; optional `domain` filter
- `get_project_prompts` - tracked prompts, visibility status, and ranks; optional `status`
- `get_project_snippets` - how AI describes a brand
- `get_project_competitors` - tracked competitor brands
## Workflow
```
- [ ] get_project_competitors
- [ ] get_project_prompts
- [ ] get_project_snippets (your brand = projectSlug)
- [ ] For top 3 rivals: get_project_snippets with brand = competitor slug
```
## Output
| Prompt | Your status | Top competitor | Their rank | Gap |
| --- | --- | --- | --- | --- |
| ... | ... | ... | ... | ... |
Then:
1. **Positioning gaps** — what rivals say that you do not (from snippets)
2. **Source gaps** — domains citing them more (filter sources by competitor mentions if visible)
3. **Quick wins** — opportunity prompts where you are close (#4–7)
4. **Strategic bets** — missing prompts worth new content
No invented competitors. Only brands from `get_project_competitors`.
Related skills
To connect MCP, open your project in AI SEO Tracker and go to Settings -> Agent.