Model-facing capability harness for LLM agent workflows: claim verification, memory continuity, risk routing, bounded improvement, and adapter contracts. DOI: 10.5281/zenodo.21189879.
powershell codex github-actions ai-agent conversation-memory doubao claim-verification agent-memory claude-code coding-agent agent-governance agent-harness memory-isolation workbuddy claim-boundary adapter-framework risk-routing evidence-gating meta-first conversation-ledger
-
Updated
Jul 25, 2026 - Python