Skip to content

Latest commit

 

History

History
43 lines (32 loc) · 2.08 KB

File metadata and controls

43 lines (32 loc) · 2.08 KB

Reputation System

TAO Passport uses weighted reputation signals to calculate wallet trust.

Starter Signal Groups

Signal Source Purpose
Validator reliability Chain Measures validator consistency and operational trust.
Miner participation Chain Rewards durable subnet mining activity.
Governance activity Chain Identifies users participating in protocol decisions.
Subnet participation Chain Rewards sustained activity across Bittensor subnet roles and economies.
Community signal Community Captures durable non-code ecosystem contribution.

Scoring Principles

  • Scores must be explainable.
  • Weights must be documented.
  • Signals should prefer durable behavior over one-time actions.
  • Abuse risks should be reviewed before increasing a signal weight.

Provenance Metadata

Each reputation signal and timeline event should carry provenance metadata so a score can be audited instead of treated as a black box.

Provenance Fields

Field Purpose
sourceCategory Classifies whether the evidence came from chain, GitTensor, community, or a derived aggregate.
sourceId Stable identifier for the underlying source record or aggregate.
reference Human-readable reference used during debugging and support review.
observedAt Timestamp for when the source evidence was last observed or refreshed.
scoringModelVersion Reputation model version used when the score component was produced.
confidence Maintainer confidence in the evidence quality (high, medium, low).
evidenceLinks Optional links to methodology or source material relevant to the signal.

Interpretation

  • Chain and derived signals should usually have high confidence when sourced from fresh indexed data.
  • Community signals may have lower confidence because they often require softer heuristics or manual review.
  • observedAt should be shown with cache freshness metadata so stale evidence is visible.
  • scoringModelVersion should change whenever scoring weights or source interpretation changes.