Skip to main content

Team Intelligence Engine

Vartovii Team Intelligence is an automated system designed to evaluate the quality, credibility, and safety of development teams behind crypto projects, specifically focusing on Pre-Launch and Tokenless stages where on-chain data is limited.

Core Purpose​

To answer three fundamental questions about any project team:

  1. Capability: Can they build what they promise? (Code Activity)
  2. Credibility: Who are they? (Background & Reputation)
  3. Safety: Are they scammers? (Risk & Rug Pull Detection)

Analysis Methodology​

The engine aggregates data from GitHub (primary source) and cross-references it with internal reputation databases.

1. The Team Score (0-100)​

The final Team Score is a composite metric derived from multiple factors. It starts at a base of 50 and is adjusted based on positive signals (Boosts) and negative signals (Penalties).

A. Technical Capability (Activity)​

  • Team Size:
    • 50+ contributors: +20 points (Enterprise scale)
    • 20+ contributors: +10 points (Strong startup)
    • 5+ contributors: +5 points (Small team)
    • < 3 contributors: -20 points (High risk, "One man army")
  • Contribution Intensity:
    • Top contributor > 500 commits: +10 points (Highly dedicated lead)
    • Top contributor > 100 commits: +5 points

B. Reputation & Background (Credibility)​

  • Tier 1 Background Check (+20 points):
    • The system scans bio/company fields of top 5 contributors.
    • Matches: Google, Meta, Apple, Coinbase, OpenSea, Paradigm, a16z, Jump Crypto, Citadel, etc.
    • Logic: Ex-FAANG or Ex-VC founders significantly increase project success probability.
  • Organization Credibility:

    • 10k Followers: +10 points

    • 1k Followers: +5 points

C. Anonymity Penalty​

  • Undoxxed Team (-15 points):
    • If more than 60% of the top 5 core contributors have no real name, Twitter link, or personal blog listed on GitHub.
    • Rationale: Anonymous teams have a historically higher rate of abandonment.

D. Risk & Rug Pull Detection (Safety)​

  • Critical Risk Flag (-50 points):
    • Known Bad Actors: Checks usernames against an internal blacklist of known scammers/rug-pullers.
    • Suspicious Signals: Scans user bios for keywords associated with low-quality "pump and dump" schemes (e.g., "pump fun", "safemoon killer", "ponzi").
    • Impact: A flag here essentially destroys the score, marking the project as "Critical Risk".

Technical Implementation​

Service Architecture​

  • Module: vartovii.modules.crypto.services.team_analyzer
  • Class: TeamAnalyzer

Data Flow​

  1. Discovery: Identify the correct GitHub Organization (e.g., solana-labs).
  2. Extraction: Fetch public repos, follower counts, and iterate through contributors.
  3. Core Team Identification: Sort by contributions to isolate the "Core 5" developers.
  4. Deep Scan: Fetch detailed user profiles for the Core Team (Bio, Location, Company).
  5. Scoring: Apply the algorithm described above.
  6. Persistence: Save team_score, team_size, core_contributors (JSON), and reputation_flags to PostgreSQL.