competitive-intelligence

name	competitive-intelligence
description	Monitor competitors, track market trends, analyze sentiment, and deliver competitive briefings. Use when the user asks about competitors, competitive analysis, market research, industry trends, competitor pricing, or market monitoring.

Competitive Intelligence & Market Research

name: competitive-intelligence
version: 1.0.0
updated: 2026-02-22
author: InstaClaw
triggers:
  keywords: [competitor, competitive analysis, market research, industry analysis, competitor pricing, market trends, intel, surveillance, monitoring, sentiment]
  phrases: ["monitor competitors", "track the market", "competitive landscape", "what are competitors doing", "market research report", "industry trends", "watch competitor pricing", "crypto sentiment", "track mentions", "who are my competitors"]
  NOT: [internal analytics, our own metrics, user analytics, A/B test results]

Overview

You are an always-on competitive intelligence analyst. You monitor competitors, track market trends, analyze sentiment, and deliver actionable briefings — daily digests, weekly deep-dives, and real-time alerts.

Primary tool: Brave Search API (required for 80% of this skill). Secondary tools: web_fetch (full page reads), file system (historical snapshots).

Without Brave Search API configured, this skill is ~20% functional (limited to fetching known URLs).

Prerequisites (on your VM)

• BRAVE_SEARCH_API_KEY in ~/.openclaw/.env (platform-provided)
• Helper scripts: ~/scripts/competitive-intel.sh, ~/scripts/competitive-intel.py
• Workspace directory: ~/.openclaw/workspace/competitive-intel/
• Heartbeat integration for scheduled scans

Helper Scripts

competitive-intel.sh — Brave Search API Client

# Web search
~/scripts/competitive-intel.sh search --query "CompetitorX announcement" --count 10
~/scripts/competitive-intel.sh search --query "site:competitor.com/blog" --freshness pd

# News search
~/scripts/competitive-intel.sh news --query "CompetitorX funding" --count 5

# Fetch and snapshot a page
~/scripts/competitive-intel.sh snapshot --url "https://competitor.com/pricing" --competitor CompetitorX --category pricing

# Rate limit status
~/scripts/competitive-intel.sh rate-status

competitive-intel.py — Analysis Engine

# Generate daily digest
python3 ~/scripts/competitive-intel.py digest

# Generate weekly report
python3 ~/scripts/competitive-intel.py weekly-report

# Compare snapshots
python3 ~/scripts/competitive-intel.py compare --competitor CompetitorX --category pricing

# Scan all competitors (runs all workflows)
python3 ~/scripts/competitive-intel.py scan

# Initialize monitoring for a new competitor
python3 ~/scripts/competitive-intel.py init --competitor "CompetitorX" --domain "competitorx.com"

Workflow 1: Competitor Monitoring Setup

When user says: "Monitor my competitors" or agent detects competitive context.

STEP 1: Identify Competitors

Agent: "Who are your main competitors? I'll set up daily monitoring."

If user provides names → proceed
If user unsure → Search "{user's product category} alternatives"
  and suggest top 5 competitors found

STEP 2: Build Competitor Profile For each competitor, create a config in ~/.openclaw/workspace/competitive-intel/config.json:

{
  "competitors": [
    {
      "name": "CompetitorX",
      "domain": "competitorx.com",
      "urls": {
        "pricing": "https://competitorx.com/pricing",
        "blog": "https://competitorx.com/blog",
        "changelog": "https://competitorx.com/changelog",
        "careers": "https://competitorx.com/careers"
      },
      "search_queries": [
        "\"CompetitorX\" announcement",
        "site:twitter.com \"CompetitorX\"",
        "\"CompetitorX\" funding OR raised OR series"
      ],
      "social_handles": {
        "twitter": "@competitorx"
      },
      "priority": 
    
  
   
     
     
     
     
     
  

STEP 3: Create Baseline Snapshot Fetch all current data and store as day-zero:

~/.openclaw/workspace/competitive-intel/
  config.json
  snapshots/
    2026-02-22-competitorx.json
    2026-02-22-competitory.json
  reports/
    daily/
    weekly/

Workflow 2: Daily Competitive Digest

Runs automatically every morning via heartbeat.

STEP 1: Price Check (2-3 API calls per competitor)
├── Fetch current pricing pages
├── Compare to stored snapshots
└── Flag any changes

STEP 2: Content Scan (1-2 API calls per competitor)
├── Search site:competitor.com/blog OR /changelog
├── Filter for new posts since last check
└── Summarize new content

STEP 3: Social Scan (2-3 API calls total)
├── Search for competitor mentions across platforms
├── Classify sentiment (positive/negative/neutral)
└── Identify trending complaints or praise

STEP 4: Job Scan (weekly only, 1-2 API calls per competitor)
├── Search LinkedIn/Indeed for company name
├── Compare to previous week
└── Analyze hiring signals (growth, expansion, pivots)

STEP 5: Assemble & Deliver

Example daily digest:

🔍 Daily Competitive Intel — Feb 22, 2026

🚨 URGENT
• CompetitorX raised Series B ($50M) — TechCrunch, 2h ago
• CompetitorY launched AI voice feature — their blog, 5h ago

💰 PRICING
• No changes detected across 3 competitors

📢 MENTIONS (24h)
• CompetitorX: 47 mentions (+12% vs yesterday)
  - Positive: 68% | Negative: 17% | Neutral: 15%
  - Top complaint: "Still no mobile app"

📝 CONTENT
• CompetitorZ published: "How to Build AI Agents" (est. 4k words)

⏱ Your scan time: 2 minutes

Workflow 3: Weekly Deep-Dive Report

More strategic analysis, delivered Sunday evening:

• Executive summary (top 3 developments)
• Per-competitor deep-dives (funding, hiring, pricing, mentions, sentiment)
• Market trends (search volume changes, content themes)
• Pricing matrix comparison
• Strategic recommendations

Workflow 4: Real-Time Alerts

Triggered immediately for critical changes:

Critical (notify immediately):

• Competitor funding announcement
• Major feature launch
• Significant price change (>10%)
• Competitor acquisition or merger
• Negative sentiment spike (>50%)
• Your company mentioned alongside competitor

Informational (bundle into daily digest):

• New blog post
• Minor price change
• New job posting
• App store review

Alert format:

🚨 ALERT: CompetitorX Funding

CompetitorX raised $50M Series B
Source: TechCrunch (15 min ago)

Implications:
• War chest for customer acquisition
• Likely pricing pressure in 3-6 months

Suggested actions:
• Review pricing strategy
• Lock in annual contracts with key customers

Workflow 5: Crypto-Specific Intelligence

For users with crypto/web3 interests (detected from USER.md):

What works with Brave Search:

• Project announcements (Twitter/Medium/blogs) — 9/10
• Crypto Twitter sentiment — 8/10
• GitHub commit activity — 7/10 (via GitHub API)
• Partnership announcements — 8/10
• Conference appearances — 7/10

What needs specialized APIs (not search):

Source	Feasibility	Better Tool
Token price movements	5/10	CoinGecko API (free)
Whale wallet activity	2/10	Etherscan/Nansen API
DEX volume changes	4/10	DexScreener API
On-chain metrics	3/10	The Graph, Dune Analytics

Recommended approach: Hybrid — Brave Search for announcements/sentiment/content + crypto-specific APIs (CoinGecko, GitHub) for real-time data. Agent should have both capabilities.

Crypto sentiment keywords:

• Positive: "moon", "gem", "bullish", "buying", "accumulating", "undervalued", "based"
• Negative: "rug", "scam", "bearish", "selling", "dumping", "dead", "overvalued"

Data Source Feasibility Matrix

Intel Category	Feasibility	Method	Reliability
Competitor pricing changes	8/10	Fetch pricing pages, compare snapshots	HIGH
New feature launches	9/10	Search site:competitor.com/changelog	VERY HIGH
Job postings (hiring signals)	7/10	Search LinkedIn/Indeed	MEDIUM-HIGH
Social media mentions	8/10	Search site:twitter.com "CompanyName"	HIGH
Funding rounds	9/10	Search Crunchbase/TechCrunch	VERY HIGH
Content publishing frequency	9/10	Search site:competitor.com/blog	VERY HIGH
App Store reviews	6/10	Product Hunt via search	MEDIUM
SEO keyword rankings	5/10	Directional only, not precise	MEDIUM
Website traffic estimates	3/10	Requires SimilarWeb/Ahrefs	LOW

Data Storage

Snapshot format (snapshots/YYYY-MM-DD-competitorname.json):

{
  "date": "2026-02-22",
  "competitor": "CompetitorX",
  "pricing": { "starter": 29, "pro": 99, "enterprise": "custom" },
  "social": { "twitter_mentions_7d": 89 },
  "content": { "blog_posts_30d": 8, "changelog_updates_30d": 3 },
  "hiring": { "open_positions": 7, "new_this_week": 2 }
}

Comparison logic: Load today's snapshot vs last week's snapshot, compute deltas for pricing, social, content, hiring. Report percentage changes.

Phase 2: SQLite Database (If Scaling)

When snapshot count exceeds 100 files, migrate to SQLite for efficient querying:

CREATE TABLE competitor_snapshots (
  id INTEGER PRIMARY KEY,
  date DATE,
  competitor TEXT,
  category TEXT,  -- pricing, social, hiring, features
  data JSON,
  created_at TIMESTAMP
);

CREATE TABLE price_changes (
  id INTEGER PRIMARY KEY,
  date DATE,
  competitor TEXT,
  tier TEXT,
  old_price REAL,
  new_price REAL,
  change_pct REAL
);

-- "Show all CompetitorX price changes in Q1"
SELECT * FROM price_changes WHERE competitor = 'CompetitorX' AND date >= '2026-01-01';

-- "Average blog posts per month by competitor"
SELECT competitor, AVG(json_extract(data, '$.blog_posts_30d'))
FROM competitor_snapshots WHERE category = 'content' GROUP BY competitor;

Migration path: Start with JSON files → SQLite when data grows. Agent handles migration automatically when it detects >100 snapshot files in the snapshots directory.

Rate Limiting & Budget

Brave Search calls:    30 per competitor/day (5 competitors = 150 total)
Web fetch calls:       20 per day
Total daily limit:     200 (hard cap)

Allocation per scan type:
  Price check:    3 per competitor
  Content scan:   2 per competitor
  Social scan:    3 total (batch queries)
  Job scan:       2 per competitor (weekly only)
  Alerts reserve: 10 for alert follow-ups

Common Mistakes

1. Over-alerting — Only critical events (funding, major launches, price changes >10%) get real-time alerts. Everything else goes in the daily digest.
2. Stale comparisons — Always show the date of the last snapshot when reporting changes.
3. Confusing correlation with causation — Present data, don't speculate on connections unless evidence supports it.
4. Ignoring rate limits — 5 competitors x 10 queries = 50 API calls. Budget carefully. Job scans are weekly, not daily.
5. Presenting search rankings as precise — Say "approximately rank #8" not "rank #8."

Quality Checklist

• All data includes source URL and timestamp
• Price comparisons show old + new values with percentage change
• Sentiment analysis includes sample size (not just percentages)
• Weekly report includes strategic recommendations
• Real-time alerts include "so what" implications and suggested actions
• Historical snapshots stored after every scan
• Rate limits respected — total API calls within daily budget
• Competitor data is accurate (no hallucinated company details)
• Delivery format matches user preference (Telegram/Discord/email)

---

Error Handling & Recovery

Tool Failure Protocol

If ANY tool or script fails during competitive intelligence gathering:

1. NEVER go silent. The user is waiting for your response.
2. Report what you were able to gather before the failure.
3. If competitive-intel.sh or competitive-intel.py fails, fall back to web_search + web_fetch manually.
4. If the Brave Search API is down, tell the user and suggest trying again later.

Heartbeat Digest Failures

• If the daily/weekly digest fails during a heartbeat cycle, log the failure to MEMORY.md.
• On the next user interaction, briefly mention: "Your latest competitive digest couldn't be generated due to [reason]. I'll retry on the next cycle."
• NEVER silently skip a scheduled digest without logging.

Timeout Handling

• If any API call takes more than 30 seconds, abort it and report partial results.
• If competitive-intel.py analysis takes more than 120 seconds, kill it and deliver whatever was gathered.
• NEVER wait indefinitely for a script to complete.

Retry Limits

• Maximum 2 retries per data source.
• If a source fails twice, skip it and note it as "[source] unavailable" in the report.
• NEVER enter a retry loop. NEVER go silent after an error.