Competitive intel

Competitive Intel

Competitor monitoring agent. Visits your configured competitor websites, detects changes in pricing, features, messaging, and positioning, and returns a structured alert with recommended actions.

What this skill does

1. Fetches up to 5 competitor websites from CONFIG.md
2. Scans pricing pages, feature pages, and homepages
3. Uses Claude to identify meaningful changes vs noise
4. Returns a structured report with alert level and recommended response
5. Flags only changes that matter — ignores minor copy edits

Configuration

Before running, set in CONFIG.md:

COMPETITOR_1=              # Company name | website URL | pages to monitor
COMPETITOR_2=              # e.g. Acme Corp | acme.com | /pricing, /features
COMPETITOR_3=
COMPETITOR_4=
COMPETITOR_5=
YOUR_ICP_INDUSTRY=
YOUR_COMPANY=
ANTHROPIC_API_KEY=

Pages worth monitoring per competitor:

• /pricing — pricing changes, new tiers, removed plans
• /features — new capabilities, renamed features
• / (homepage) — messaging and positioning shifts
• /blog — content strategy signals
• /careers — hiring signals (what they are building next)

How to invoke

# CLI
agent run competitive-intel

# Voice (after ACTIVATION.md setup)
"[your wake word] competitors"

# Scheduled — set your preferred day and time
# Recommended: weekly, not daily — daily creates noise
# See ACTIVATION.md for scheduling instructions

Alert levels

Level	Meaning	What triggers it
🔴 RED	Act immediately	Pricing cut, major new feature targeting your ICP, direct positioning attack
🟡 YELLOW	Monitor closely	New blog content, minor feature update, job postings in key areas
🟢 GREEN	No action needed	No meaningful changes detected

Claude system prompt (what runs under the hood)

You are a competitive intelligence analyst for a company in the [YOUR_ICP_INDUSTRY] space.

Review the following competitor website content and identify any meaningful changes 
in: pricing, features, positioning, messaging, or target audience.

Ignore: minor copy edits, blog posts about general topics, cosmetic design changes.

Flag: pricing changes, new features, new target segments, repositioning, 
aggressive messaging directed at companies like [YOUR_COMPANY].

For each competitor return:
- competitor_name
- alert_level (RED / YELLOW / GREEN)
- changes_detected: list of specific changes found
- what_it_means: 2 sentences on the strategic implication
- recommended_response: one specific action to take or consider
- pages_checked: list of pages reviewed

Also return an overall_alert_level for the full report.
Return valid JSON only.

Output format

{
  "overall_alert_level": "YELLOW",
  "competitors": [
    {
      "competitor_name": "Competitor name",
      "alert_level": "YELLOW",
      "changes_detected": [
        "Added a new free tier on pricing page",
        "Homepage headline changed from X to Y"
      ],
      "what_it_means": "They are moving down-market with a freemium play. This could attract your ICP at the top of funnel before you reach them.",
      "recommended_response": "Review your own free trial or lead magnet offer — ensure your value is clear before they hit a competitor paywall.",
      "pages_checked": ["/pricing", "/", "/features"]
    }
  ],
  "report_date": "2026-06-16",
  "next_recommended_run": "7 days"
}

Chaining with other skills

# Turn competitor changes into a LinkedIn post
agent run competitive-intel | agent run linkedin-post-generator

# Natural language
agent ask "Check competitors and write a LinkedIn post about what you found"

Limitations

• This skill reads publicly visible web pages only
• It cannot access paywalled content, logged-in areas, or private data
• Accuracy depends on how frequently you run it — weekly is recommended
• Some websites block automated requests — add backup monitoring manually if needed

Related skills

• linkedin-post-generator — turn competitor observations into content
• ai-news-digest — combine with news for a full market picture
• See ACTIVATION.md for voice and scheduled setup