Issues

• Chassis Shootout G5 — paired token/cost/complexity analysis and chassis decision

  oigorbrito/codepro#103 · Filter by author oigorbrito opened on Oct 7, 2026

• Chassis Shootout G4 — run unmodified Agentless baseline

  oigorbrito/codepro#102 · Filter by author oigorbrito opened on Oct 7, 2026

• Chassis Shootout G3 — run unmodified nano-harness baseline

  oigorbrito/codepro#101 · Filter by author oigorbrito opened on Oct 7, 2026

• Chassis Shootout G2 — freeze workload, verifier, budgets, and loop detector

  oigorbrito/codepro#100 · Filter by author oigorbrito opened on Oct 7, 2026

• Chassis Shootout v1 — Nano vs Agentless under identical evidence controls

  oigorbrito/codepro#98 · Filter by author oigorbrito opened on Oct 7, 2026

• Evaluate subscription-backed CLI executor adapter mechanisms

  oigorbrito/codepro#97 · Filter by author oigorbrito opened on Oct 6, 2026
  

• research: empirically test decomposition, context, routing, coordination, and token goodput

  oigorbrito/codepro#93 · Filter by author oigorbrito opened on Oct 3, 2026

• Architecture Confrontation v1 — benchmark external scaffolds against CodePro hot-path complexity

  oigorbrito/codepro#84 · Filter by author oigorbrito opened on Sep 28, 2026
  1

• Benchmark Decision Protocol v1 — paired evidence and scaffold provenance

  oigorbrito/codepro#78 · Filter by author oigorbrito opened on Sep 26, 2026
  1

• Harness Change Regularization v1 — test held-out transfer and cost-aware acceptance

  oigorbrito/codepro#37 · Filter by author oigorbrito opened on Sep 24, 2026
  2

• Benchmark Expansion v1 — assess SWE-bench Science / Hard70 for CodePro evaluation

  oigorbrito/codepro#35 · Filter by author oigorbrito opened on Sep 24, 2026
  2

• Selective Retrieval v1 — test grounded context injection versus anchoring failure

  oigorbrito/codepro#34 · Filter by author oigorbrito opened on Sep 24, 2026
  2

• Context Memory v1 — compare searchable cross-window state against lossy compaction

  oigorbrito/codepro#33 · Filter by author oigorbrito opened on Sep 24, 2026
  1

• Executor Cost-Capability v1 — reproduce new coding-agent efficiency results under CodePro controls

  oigorbrito/codepro#32 · Filter by author oigorbrito opened on Sep 24, 2026
  2

• Mini Efficiency v1 — optimize mini-swe-agent without changing executor semantics

  oigorbrito/codepro#31 · Filter by author oigorbrito opened on Sep 24, 2026