Cotana Hack Tracker

Detection Methodology

Evidence-first behavior scoring for incident classification and confidence evolution.

What Cotana Detects

  • Wallet drainer behavior and abnormal outflow patterns.
  • Protocol exploit signatures from contract balance deltas and method anomalies.
  • Bridge exploit markers including unbacked mint and liquidity cliff behavior.
  • LP exploits: exploits that drain value directly from a liquidity pool or AMM by manipulating pool mechanics, balances, or withdrawal paths.

What Cotana Does Not Detect

  • No private off-chain intelligence or non-public incident feeds.
  • No protocol-specific hardcoded exploit labels.
  • No manual human narrative injection in generated summaries.
  • No certainty guarantees; confidence depends on visible on-chain evidence quality.

Confidence Scoring

Confidence is behavior-derived from independent rule hits, interaction weighting between correlated vectors, and lifecycle updates over time. Scores rise when multiple high-signal vectors align in the same window and actor graph.

Summaries and timelines are generated deterministically from stored incident data and link directly to supporting on-chain evidence.