llm-agents

DeepSeek Harness 的执行迷宫——看 Agent 真实怎么干活：迷宫时间轴 · 数据轨道 · 确定性执行分析 · 多会话对比 | The execution maze for DSH agents: maze timeline, per-step data tracks, deterministic execution analysis, multi-session comparison. Formerly dsh-trace-compare.

DSH plugin — free OpenCode Zen models for DeepSeek Harness (DSH). Free LLM API, no API key needed. 在 DSH 中使用 OpenCode Zen 免费模型，无需 API key

专为 DeepSeek Harness Web GUI 打造的可视化多 Agent 工作流编排插件。公开测试版已上线！接下来会随着 DSH 的正式版上线一并发布正式版！最新声明：非常抱歉，经过昨日的一个测试，发现编排引擎会偶发提示词交叉，导致子代理越权行为，并且自己调用已限制工具，可能会有安全隐患，正在修复。

DSH Math Modeling Team Plugin Pack: 2 role agent presets (modeling-coding / paper) for DeepSeek Harness, multi-folder Git collaboration + vision sub-agent; includes a full run on 2023 CUMCM-C. ｜ DeepSeek Harness 数学建模团队插件包：2 套岗位Agent预设(建模编程+论文)、多文件夹协同+识图子代理，含2023国赛C题全流程跑通示例。

X-ray for DeepSeek Harness plugins: declared capabilities vs actual behavior. Registry + static scanner + badges.

Project Memory gives coding agents a single, trustworthy memory for a software repository — so they stop re-learning the same facts and stop writing conflicting "memory" files.

DeepSeek Harness (DSH) 代码参考检索与复用调查插件：开发前检索本地代码库与 GitHub/npm 开源项目，评估复用/改造/引入依赖/自制并询问用户，支持架构级复用。Code reference & reuse-survey plugin for DSH: search local code + GitHub/npm before building, evaluate reuse vs rewrite, ask the user, architecture-level reuse.

Explain failed DeepSeek Harness agent runs locally, then produce a redacted recovery plan. No model required.

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.

该仓库暂未提供项目说明。

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

AutoResearch 科研协议 DSH 全家桶插件：arXiv recall + LQS 评分 + 状态持久化 + 停滞红绿灯 + 五人格中位数评审 + 可视化总控台卡片。灵感：Deli Chen "From Draft to Strong-Accept: How a Self-Play Survey Hit 8.6"。

DeepSeek Harness 批量并行子 Agent 调度插件：swarm_batch 工具（Kimi Code Swarm 模式移植）+ /swarm 命令 + 聊天内实时进度条

Interactive multi-agent collaboration, meetings, group chats, and task execution for DeepSeek Harness.

zenmux for deepseek harness

该仓库暂未提供项目说明。

Loop guard for DeepSeek Harness: denies repeated identical tool calls and caps per-agent calls, outside the model where instructions cannot reach

Event-based long-term memory for LLM agents — associative surfacing over query RAG. Closed-loop episodic events, budgeted injection, tiered forgetting. Claude Code hooks + DeepSeek Harness plugin, one shared .memory/ format.

Workflow-net runtime for DeepSeek Harness: resource-aware concurrency, native loops and fan-out, static soundness checking before a plan becomes durable, and process mining over its own event log.

Context firewall for DeepSeek Harness: credentials in tool results become opaque placeholders before they reach the model — the secret never enters context.

该仓库暂未提供项目说明。

Long tasks shouldn't cost like they need a frontier model. Cheap-model long-task engine: reliability in structure (disk state machine + mechanical gates + T3 zero-prose dispatch + multi-brain audit), not in model strength. Pure Python stdlib. DSH skill.

Forgetting engine for AI agents — memory TTL, decay, eviction, audit. The opposite of memory programming. DSH plugin + local skill + CLI.

dsh-preset-run: DeepSeek Harness plugin — preset_run tool runs one-shot tasks under any agent preset (router-spec/router-standard/minimal) headlessly. Install with dsh plugin add.

给 DSH agent 的插件库"起飞前检查单"——任务型请求时自动预检插件库并注入 Top-K 适用插件提示，让插件库利用率可预期、不靠运气。Pre-flight plugin-library check for DSH agents — task-type requests trigger a Top-K hint of suitable plugins injected into the runtime context, making plugin usage predictable instead of opportunistic.

生信模式 (Bioinformatics Mode) — research-grade protein structure & interaction agent preset for DeepSeek Harness: AF2-Multimer, ESMFold, AutoDock Vina, MM-GBSA/MD, structure QA, virtual screening.