Deep dive

System Architecture

Understanding the core mechanics of JudgeNod's verifiable scoring engine and its integration with the Solana blockchain.

High-level architecture of the JudgeNod smart contract ecosystem.
— High-level architecture of the JudgeNod smart contract ecosystem.

The Scoring Engine

JudgeNod's scoring engine is split into two layers: a performant off-chain analysis layer and an atomic on-chain verification layer.

1. Off-chain Analysis (The CLI)

The CLI tool performs a deep analysis of target repositories, including:

  • Code quality metrics (complexity, documentation coverage).
  • Test suite execution and coverage reports.
  • Commit history and contributor attribution.

2. On-chain Verification (The Smart Contract)

The Solana smart contract (built with Anchor) acts as the source of truth. It stores:

judge_contract/lib.rs
#[account] pub struct Rubric { pub organizer: Pubkey, pub criteria: Vec<Criterion>, pub total_weight: u8, pub is_finalized: bool, }  #[account] pub struct Submission { pub judge: Pubkey, pub project_hash: [u8; 32], pub scores: Vec<u8>, pub timestamp: i64, }
The contract uses Ed25519 signature verification to ensure that only authorized judges can submit scores for a specific rubric.

Data Flow

When a judge evaluates a project, the following sequence occurs:

  1. Ingestion: The project repository is cloned and analyzed locally.
  2. Hashing: A unique cryptographic hash of the analyzed code state is generated.
  3. Submission: The score, along with the project hash, is sent to the Solana program.
  4. Finalization: Once the judging period ends, the organizer triggers a finalization transaction that computes the final immutable leaderboard.
This architecture ensures that the leaderboard can be computationally reproduced by anyone with access to the original repositories and the on-chain submission records.
Last updated: April 16, 2026
Edit this page on GitHub &nearr;