Song Luo

RedinGhost

1

https://github.com/rrrrrredy

AI & ML interests

I am an AI Agent researcher and engineer building practical, evidence-driven harnesses for agents: the runtime, tool-use, memory, web access, skill execution, context packaging, safety, observability, and evaluation layers that turn agents from demos into working systems.

Recent Activity

updated a dataset 12 days ago

RedinGhost/DeliveryBench-ZH

updated a dataset 12 days ago

RedinGhost/StyleBench-ZH

updated a Space about 2 months ago

RedinGhost/editorial-guard

View all activity

Organizations

None yet

Collections 2

spaces 12

Editorial Guard

Chinese editing tools, benchmarks and full source repository

Research Toolkit

Evidence-led research, effective review and delivery checks

Agent Memory System

Evidence-backed continuous memory for coding agents

BeforeDone

Fresh verification evidence before coding agents say done

Execution Fidelity Guard

Task-contract gates and evidence-aware completion checks

Context Continuity

Preserve task invariants across compaction and handoff

models 0

None public yet

datasets 7

RedinGhost/DeliveryBench-ZH

Viewer • Updated 12 days ago • 482 • 84

RedinGhost/StyleBench-ZH

Viewer • Updated 12 days ago • 482 • 105

RedinGhost/research-toolkit-evals

Viewer • Updated Sep 21 • 34 • 47

RedinGhost/beforedone-paper

Viewer • Updated Sep 2 • 169 • 85

RedinGhost/governed-skill-evolution

Viewer • Updated Sep 2 • 48 • 44

RedinGhost/skillops-paper

Viewer • Updated Jul 28 • 101 • 89

RedinGhost/agent-harness-paper

Viewer • Updated Jul 28 • 240 • 40