benchmark

GLM-5.3-Flash × J-Space capability realization — benchmark presentation of the J-Space Cognition Suite

Design-fidelity QA for DeepSeek Harness: lend any text-only model an eye, then judge whether the implementation matches the mock. Ships the benchmark behind that judgement — four fixtures, 23 injected defects, and every raw model transcript. Retires itself when DeepSeek ships vision.

dsh-excel-chat — talk to Excel in DeepSeek Harness: create, edit, repair, and verify spreadsheets by conversation (cells, formulas, styles, filters, tables, charts); every edit is auto-validated.

Execution-time drift firewall for long-running DeepSeek Harness agents. Real-Harness tests: unsafe stale mutations 12/12 native -> 0/12; valid controls 7/7 both; post-SIGKILL unsafe continuation 2/2 -> 0/2.

Make your research agent hard to fool — the discipline layer for agentic research: breakthrough loop, hard-to-fool harness checklist, claim-polarity gates, engineering rules with their tuition

DeepSeek Harness tools for reproducing the ml-quant-trading protocol v1 benchmark.

Deterministic revision-pinned benchmarks and regression evidence for DeepSeek Harness

知设 DSH 插件共享基础设施 - 知识库加载/检索/基准价格/风险评估

Agent evaluation platform for DeepSeek Harness: benchmark YAML, headless dsh orchestration, trace-based metrics, LLM judge, paired A/B, keyless replay, and cross-harness import.

Evidence-backed, type-aware quality scorecards for DeepSeek Harness plugins.

该仓库暂未提供项目说明。

A blind, fair, local DSH Web arena: same task, isolated worktrees, shared verification, judge before reveal.

Host and Web performance diagnostics, repeated plugin-isolation campaigns, and safe recovery for DeepSeek Harness.

该仓库暂未提供项目说明。