Competitive Intelligence

Why this skill

Competitive analysis tends to scatter across notes, battle cards, and half-remembered opinions that never connect back to how you win deals. This skill produces one structured analysis anchored to your value architecture, mapping each competitor to where you should compete, co-exist, or step aside.

What it does

The skill runs in four phases:

• Discover: find competitors and research each across ten dimensions, optionally through a deep internet search that surfaces players you do not know yet.
• Categorize: assign each competitor a type and a threat level, and map where it competes with you.
• Score: rate each competitor against your value drivers to find your moat and your gaps.
• Synthesize: write a profile per competitor and a summary per category.

The Discover phase researches each competitor across ten dimensions:

1. Company profile: founding, location, team size, funding, ownership.
2. Product: positioning, features, differentiators, integrations, delivery model.
3. Customer fit: target customer, company sizes, industries, personas, geography, regulatory context, named customers.
4. Commercial terms: pricing model, tiers, value metric, free and enterprise availability.
5. Buyer sentiment: G2 and Capterra ratings, what reviewers praise and criticize, community signals.
6. Content and momentum: blog cadence, topics, formats, recent launches.
7. Sales motion: sales-led, product-led, or hybrid, primary call to action, outbound signals.
8. Hiring signals: open roles by department and what the hiring pattern implies.
9. Digital footprint: LinkedIn presence and cadence, search rankings, indexed page count.
10. Advertising signals: active LinkedIn, Google, and Meta ads and their creative themes.

The output lands in context/competition/: one profile file per competitor, one summary per category, and two HTML visuals, a competitor overview and a differentiation matrix. Every claim carries its source, and anything that is not publicly findable is flagged as a data gap rather than guessed.

How to use it

Call /act-competitive-intelligence. It asks a short set of setup questions up front, including output language, whether to run a deep search or work only from competitors you name, and the company name for the visuals. Then it runs on its own.

When it finishes it reports the files it wrote, a threat summary by category, your win and avoid zones by customer profile, a short watch list of competitors to track, and any data gaps to fill. The result feeds directly into the website Plan step, which reads context/competition/ when it builds comparison pages. It does not commit or publish anything.

Optional research tools

The skill runs on built-in web search by default. Two optional integrations sharpen the last two dimensions when they are available, and the skill falls back on its own when they are not.

• Serper.dev, a search API: add a SERPER_API_KEY to your project's .env.local and the skill uses it for the digital footprint dimension, pulling exact search rankings, featured-snippet ownership, and indexed page counts. Without a key it uses standard web search and records which method produced each result. The key is read at call time and never printed or written into output.
• Apify, connected as an MCP server: with it, the skill reads advertising data from the Google, LinkedIn, and Meta ad libraries, which are otherwise hard to read because those pages render with JavaScript. Without it, the skill notes that a manual check is needed rather than guessing.

For both, the free tier is enough. A single competitive analysis stays well within the monthly free-tier limits on Serper.dev and Apify, so you can run the skill without a paid subscription.