Cheggin/competitor-research

Competitor Research

Deep competitor analysis using browser agents to visit real websites. Every data point comes from observed behavior, not search result summaries.

Inputs

The user must provide:

• Target market — specific product category (e.g. "AI coding assistants", "developer analytics platforms")
• Our product (optional) — what we're building, so the report can identify where we win/lose
• Moat axes (optional, defaults to: user experience, agent experience, pricing)

Phase 1: Discovery

1. Use web search to find 5-15 competitors with verifiable URLs.
2. For each competitor, record: name, URL, one-line description, funding stage if findable.
3. Prioritize direct competitors over tangential ones. Cut anything that doesn't actually compete.

Phase 2: Browser Deep-Dive

For EACH competitor, spawn a browser agent (use Firecrawl or Playwright) to visit their actual site. The agent must visit and capture:

Pages to visit:

• Homepage — first impression, value prop clarity, CTA placement
• Pricing page — exact tiers, prices, feature gating, free tier limits, enterprise contact-us gates
• Docs/API page — SDK quality, getting-started friction, API design patterns
• Signup flow — how many steps, what info required, time-to-first-value

Data to extract per competitor:

• Screenshots of each page visited (save to screenshots/{competitor-slug}/)
• Pricing tiers — exact dollar amounts, feature lists per tier, billing model
• UX friction score (1-10) — clicks to value, signup complexity, onboarding clarity. Justify the score with specific observations.
• Agent/API experience score (1-10, or N/A) — SDK docs quality, API ergonomics, CLI UX, webhook support. Justify with specifics.
• Design quality notes — patterns worth adopting, patterns to avoid, visual polish level
• Tech stack signals — frameworks, infrastructure, analytics, payments (from page source, headers, public info)

Phase 3: Analysis

Comparison Matrix

Markdown table with competitors as columns and rows for:

• Pricing (lowest tier, mid tier, enterprise)
• Free tier limits
• UX friction score (with one-line justification)
• Agent/API experience score (with one-line justification)
• Key differentiator
• Biggest weakness

Pricing Breakdown

Detailed pricing comparison — exact numbers, not ranges. Note which competitors gate critical features behind enterprise tiers.

Moat Analysis

For each moat axis (UX, agent experience, pricing):

• Where we win
• Where we lose
• What to steal (specific patterns, not generic advice)

Positioning Gaps

Identify segments where no competitor excels. These are differentiation opportunities.

Phase 4: Report

1. Write findings to research-report.md with structured, parseable sections.
2. Every claim must link back to its source URL.
3. Every score must have a specific justification — no generic praise or criticism.
4. Run the report through /startup-harness:anti-ai-writing to strip AI slop before publishing.
5. Create a GitHub Issue with the findings, including labels and source URLs.

Constraints

• Every competitor must have a verifiable URL — never fabricate names or URLs.
• Every pricing figure must come from the actual pricing page — never guess or approximate.
• Every UX score must cite specific observed behavior — "clean design" is not a score justification.
• Screenshots are mandatory proof. If a page couldn't be captured, say so explicitly.
• This skill produces research only, not product specs or implementation plans.
• Flag the report if research is older than 30 days for staleness detection.