Brand Protection Agent

Indexes your site, simulates how AI engines retrieve from your content, scores citations, verifies factual claims, and red-teams your brand against adversarial prompts.

When to use it

• Setting up brand protection for the first time — start with /index-site
• Checking whether AI citations of your brand are factually accurate
• Stress-testing your content against adversarial prompts
• Finding semantic gaps between your content and a competitor's
• Suggesting internal links based on content similarity

How to invoke

The Brand Protection agent doesn't have an @mention — invoke its skills via the / commands below, or ask the Supervisor in natural language (for example, "simulate RAG retrieval for query X").

Skills and commands

Command	What it does
/index-site	Index or re-index a site (incremental, with PDF support)
/search-content	Semantic search across indexed brand content
/compare-content	Compare client vs. competitor content similarity
/simulate-rag	See what an LLM would retrieve from your content
/gap-analysis	Semantic gap analysis vs. competitors
/internal-links	Suggest internal links from content similarity
/citation-score	Score how well your content matches AI citations
/verify-brand	Verify factual claims in AI responses against your indexed content
/red-team	Test brand resilience against adversarial AI prompts (simulated)
/red-team-live	Send adversarial prompts to real AI engines and analyze actual responses (Ultra/Enterprise)

The agent also exposes a compare_domains skill internally for whole-domain comparisons.

Example prompts

• /index-site example.com — Crawls and indexes the site so the other skills can operate on it
• /red-team-live example.com — Sends adversarial prompts to live AI engines and reports back what happens

Related docs

• Brand Protection
• Brand Protection In Depth