All pull requests

Pull requests

• Share native evaluation execution with retained capture

  #42·PaulHax opened Oct 27, 2026Collaborator·5/5·

  ·

• Find and retrieve verified experiment evidence through MLflow

  #41·PaulHax opened Oct 27, 2026Collaborator·4/5·

  ·

• Retain checkpoint evaluation attempts and episode evidence

  #40·PaulHax opened Oct 27, 2026Collaborator·3/5·

  ·

• Match CybORG decoy exploit effects and SQL execution

  #39·PaulHax opened Oct 27, 2026Collaborator·5/5·

  ·

• Repair co-training CI checks and retain named failure diagnostics

  #38·PaulHax opened Oct 26, 2026Collaborator·3/5·

  ·

• Allow experiment-local joint environments in training and evaluation

  #36·PaulHax opened Oct 26, 2026Collaborator·4/5·

  ·

• Add versioned experiment records and explicit evidence collation

  #35·PaulHax opened Oct 26, 2026Collaborator·2/5·

  ·

• Add experiment conventions and a saved frozen-defender analysis

  #34·PaulHax opened Oct 26, 2026Collaborator·1/5·

  ·

• Add pretrained H-MARL Blue baselines and H-MARL paper Reds for evaluation

  #27·Dmujt opened Sep 28, 2026Contributor·

  ·

• Cotraining trajectory viz

  #26·PaulHax opened Sep 15, 2026Collaborator·

  ·

• Implement Red/Blue Self-Play/Co-training for CEC Experiments

  #23·Dmujt opened Aug 27, 2026Contributor·2/5·

  ·

• CC4: train-time env diversity produces ZSC to held-out red agents at 10M timesteps

  #20·PaulHax opened May 12, 2026Collaborator·1/5·6 of 7 tasks·

  ·