dsh-social-simulation
百万 agent 社交模拟:可扩展开源社媒模拟器,LLM agent 逼真模仿 Twitter/Reddit 上多达百万用户——23 种动作(follow/comment/repost)、兴趣与热度推荐算法、动态环境;研究信息传播/群体极化/羊群行为等社会现象。受 camel-ai/oasis(5k★ Apache-2.0)启发。
347 results
百万 agent 社交模拟:可扩展开源社媒模拟器,LLM agent 逼真模仿 Twitter/Reddit 上多达百万用户——23 种动作(follow/comment/repost)、兴趣与热度推荐算法、动态环境;研究信息传播/群体极化/羊群行为等社会现象。受 camel-ai/oasis(5k★ Apache-2.0)启发。
ADHD 友好输出:行动先行、步骤编号、抑制离题、每轮重申状态、具体时间估计、让进展可见、就事论事错误、列表上限 5 项、无开场白/总结/结尾语。受 ayghri/i-have-adhd(22.3k★ MIT)启发。
DSH (DeepSeek Harness) network plugin that bundles User-Agent rewriting, an HTTP/CONNECT/SOCKS5 proxy, and configurable request auto-retry for every outgoing global-fetch request, all configured from a single “网络设置” (Network) settings tab.
OpenCode Go remaining-quota widget for the DeepSeek Harness Web UI: a floating card on the right edge showing live remaining percentages for the 5h / weekly / monthly quota windows, visible only while the current model routes through OpenCode Go.
151 款 AionUi 皮肤聚合 + 皮肤中心试穿面板 / A 151-skin pack with a try-on panel for DSH Web UI.
Token optimizer for DeepSeek Harness — condense your context, keep the essence
Floating AI translation window for the DeepSeek Harness Web GUI: multilingual translation through the DSH host, with provider balance and daily-usage display. Standalone plugin package.
AI Agent runtime authorization & evidence verification — tool-call GuardrailProvider, CCS 7-dimension verification standard, MCP/DSH security scanner, SSRF/command-injection/credential-exfil blocking with Ed25519 signed receipts.
Convert PDF and Word (DOCX) documents to clean Markdown. Use when the user asks to convert PDF/Word to Markdown, extract document text/structure, 提取 PDF/Word 内容, 转 Markdown, or prepare documents for editing/RAG. Handles Chinese documents and Windows paths well.
DeepSeek Harness (dsh) LLM adapter backed by aimux — one Rust engine, 325+ providers as dsh model routes.
DSH 装备升级套件:token 费用面板(dsh-cost)、会话文件预览(dsh-plugin-file-preview)、外网搜集(dsh-research-mcp,工具名 mcp__research__*)、视觉桥接(vision-bridge,给无视觉模型看图)、微信双向通道(dsh-wechat-bridge,微信发消息→本机执行→回微信)。已安装时,用户问"花了多少钱/预览这个文件/搜一下外网/看这张图"直接按对应组件办事;未安装时给出 install.sh 一键安装命令。
附件自动降级:DSH 图片准入(单边 ≤2000px / ≤3.5MB / ≤4000万像素)超限图片不再拒绝,用 sharp 自动缩小/重编码后入库,用户上传手机原图不再踩线。
出租车计价器式实时计费:模型生成时逐秒跳动、空闲归零,内置 DeepSeek 官方峰谷时段与 V4-Flash 价格表,5 位小数。
OpenCode Go usage floating widget for the DeepSeek Harness Web UI: a bottom-right overlay showing rolling-5h / weekly / monthly quota, backed by the official opencode.ai usage endpoint.
DeepSeek Harness 极简增强模式 | Minimal-Win Agent Preset(Windows 下自动使用 pwsh 的极简增强预设)
Dual-face dsh web plugin: five-column AppFrame, right-side Git workspace tree, Source Control, and a bottom multi-terminal dock
OpenCode Go 额度悬浮窗:Web 页面右下角圆形按钮,点击展开弹窗展示 5 小时滚动/每周/每月三个周期的额度用量(percent、限额、重置时间),可随时收起为圆形按钮。
Bidirectional knowledge bridge between DeepSeek Harness and Obsidian Vault — FTS5 search, draft writing, session linking
Share any DeepSeek Harness session as a self-contained read-only static webpage — timeline, collapsible tool calls & reasoning, default redaction, dual SHA-256 fingerprints, watermark. · 把 DSH 会话一键变成可分享的只读静态网页:单文件 HTML、离线可开、默认脱敏。
DeepSeek Harness 插件:语音 + 通知出口——agent 通过云端 TTS(火山 seed-tts / 小米 MiMo V2.5,失败自动回退 SAPI)/ 桌面通知 / 提示音主动联系用户。融合 dsh-plugin-notify 的 DSH 原生深度集成与 agent-voice-mcp-minus 的云端 TTS 调优,零 Python 依赖,Windows 原生。
DSH plugin: exposes per-provider balance / coding-plan usage (5h window + weekly) in a status bar pinned to the native session-stats row.
Cross-session self-learning memory for DeepSeek Harness, ported from XT-AGENT packages/memory. BM25 relevance injection + background extraction (sanitize/dedupe/merge) + lifecycle archive + memory_read/memory_search/memory_write tools.
Read-only browser verification tools for the DeepSeek Harness web GUI: browser_open / browser_mock / browser_assert / browser_screenshot — verify a page (H5/desktop) in ≤4 tool calls with mock interception, DOM assertions, and screenshots that auto-project into the model context.
DeepSeek Harness plugin: a deterministic UTF-8 byte cap on tool results, keeping the first N bytes and carrying a sha256 of the full original so a truncated result is never mistaken for a whole one