competitor-watch

name	competitor-watch
description	Monitors competitor projects for new releases, partnerships, TVL changes, and strategic moves
tags	["research","competitive-intelligence","monitoring","strategy"]
agent	researcher
var	${var} specifies competitors to focus on. If set (e.g., "Marinade,Jito"), deep-dive those projects. If empty, scan all watched competitors.

competitor-watch

Priority: P2 (runs twice weekly) Schedule: Tuesday/Friday 10:00 UTC Data sources: GitHub API, DeFiLlama, CoinGecko, project blogs/RSS Output: Competitor intel in memory/research/competitors/

Instructions

You are executing the competitor-watch skill for the Researcher agent.

Step 1: Load Competitor List

Read memory/research/competitor-list.json:

{
  "competitors": [
    {
      "name": "ProjectX",
      "github_org": "projectx-labs",
      "token": "PROJX",
      "coingecko_id": "projectx",
      "defillama_slug": "projectx",
      "website": "https://projectx.io",
      "rss_feed": "https://blog.projectx.io/rss",
      "twitter": "projectx_defi",
      "category": "DEX"
    }
  ]
}

If the file does not exist, create it with an empty competitors array and exit with SKILL_EMPTY.

Step 2: GitHub Activity Scan

For each competitor's GitHub org:

curl -s -H "Authorization: token ${GITHUB_TOKEN}" "https://api.github.com/orgs/{github_org}/repos?sort=pushed&per_page=10"

Track:

• New repositories created since last scan
• Repos with significant recent commit activity (>10 commits in 7d)
• New releases or tags:

curl -s -H "Authorization: token ${GITHUB_TOKEN}" "https://api.github.com/repos/{org}/{repo}/releases?per_page=5"

• Open issues and PR velocity

Step 3: On-Chain and Market Data

TVL tracking:

curl -s "https://api.llama.fi/protocol/{defillama_slug}"

• Current TVL and 7d/30d change
• TVL by chain breakdown
• New chain deployments

Token performance:

curl -s "https://api.coingecko.com/api/v3/coins/{coingecko_id}?localization=false&tickers=false&community_data=true&developer_data=true"

• Price, market cap, 7d/30d change
• Developer activity score
• Community metrics (Twitter followers, Reddit subscribers)

Step 4: News and Announcements

Check for recent blog posts, announcements, partnership news:

• Fetch RSS feeds if configured
• Check for governance proposals (if DAO-governed)
• Look for new integrations or partnerships mentioned on their socials

Step 5: Comparative Analysis

For each competitor, generate a delta report:

{
  "name": "ProjectX",
  "report_date": "2024-01-15",
  "changes_since_last_scan": {
    "tvl_change_pct": 12.5,
    "tvl_current_usd": 150000000,
    "token_price_change_7d_pct": 8.3,
    "github_commits_7d": 45,
    "new_releases": ["v2.1.0 - Added limit orders"],
    "new_repos": [],
    "partnerships": ["Integrated with Wormhole"],
    "governance_proposals": [],
    "notable_events": ["Launched on Base chain"]
  },
  "threat_level": "low | medium | high",
  "threat_assessment": "ProjectX expanding to Base could pull liquidity from Solana",
  "opportunities": ["Their limit order feature suggests market demand we could capitalize on"]
}

Step 6: Save and Alert

1. Write individual reports to memory/research/competitors/{name}/{YYYY-MM-DD}.json
2. Write consolidated summary to memory/research/competitors/latest-summary.json
3. For any competitor with threat_level: "high", alert the analyst mesh
4. Send notification for significant strategic moves

Anti-Patterns

• Do NOT stalk individual developers or contributors. Focus on public project metrics.
• Do NOT make value judgments about competitor quality. Report facts.
• Do NOT access private repositories or non-public data.
• Respect GitHub API rate limits: 5000 requests/hour with token.

Exit Codes

• SKILL_OK — competitor scan complete, N competitors analyzed
• SKILL_PARTIAL — some competitors could not be fully analyzed
• SKILL_EMPTY — no significant changes detected
• SKILL_FAIL — critical failure

Output

Commit message format: researcher: competitor-watch — {N} competitors, {M} notable changes