dsh-tradingagents
TradingAgents for DeepSeek Harness: the 14-role A-share multi-agent analysis pipeline (analysts → bull/bear debate → research manager → risk debate → trader) behind the /trading-agent command. · A股多智能体投研分析插件。
1389 results
TradingAgents for DeepSeek Harness: the 14-role A-share multi-agent analysis pipeline (analysts → bull/bear debate → research manager → risk debate → trader) behind the /trading-agent command. · A股多智能体投研分析插件。
PollyReach phone-call tools and skill for DeepSeek Harness
Password-gated public access for the Web GUI via Cloudflare Tunnel: a loopback password gate (login page + HMAC cookie) in front of the webserver, published by spawning cloudflared in quick or named-tunnel mode - tells the shell and the model the public URL
Deploy the ArchGraph ARGO toolchain, skills, and rules (schema, scripts, argo-init skill, global rule) with one command.
CC-style auto mode for DeepSeek Harness: deterministic deny/allow rules + pre-execute gate + model-agnostic two-stage classifier. Two-state (allow/reject) classifier since 0.8.0. TypeScript rewrite merging dsh-auto-mode v0.4.1 with Nuo-cl/dsh-auto-mode native integration.
Fixed rail of conversation landmarks with hover previews and click-to-jump navigation for DeepSeek Harness Web
Engineering workflow layer for DeepSeek Harness (dsh): a disciplined-engineer agent preset with five gated phases — requirements clarification, plan approval, TDD implementation, parallel subagent execution, and verified finishing. Skills adapted from the obra/superpowers workflow methodology.
Use your ChatGPT subscription in DeepSeek Harness to access Codex models and generate or edit images with gpt-image-2—no OpenAI Platform API key required.
Hatch a digital creature in your DeepSeek Harness. It feeds on your real work — turns, tool calls, even your errors — and evolves along branching lines shaped by how you actually work. Zero tokens, zero tools, the model never knows.
Native continuity recovery for long-running DeepSeek Harness agents
OpenCode Go models, mixed Completions/Responses/Messages chat, and subscription usage for DeepSeek Harness
VS Code (from a code-server release) inside DSH: a right-sidebar tab driven by the plugin's own launcher over the in-process VS Code server (lib/launcher.mjs), plus an editor bridge (assets/extensions/dshcs-editor-bridge) that gives the agent what only the editor knows (unsaved buffers, language-server diagnostics, the active selection) and lets editor gestures drive the session. The bridge is completely read-only: all four of its routes (/health, /sync, /old, /event) are reads. Asking DSH from the editor opens a floating dialog inside the DSH page (draggable, resizable, bottom-right): the plugin's client half sends the question as user input (POST /api/code-server/ask/send) and renders that session's new content with DSH's own Markdown renderer - the same @deepseek-ai/dsh-client-ui-primitives instance the UI uses, required from the page's module table, so typography, highlighting and math match the interface and a renderer version mismatch is impossible. Thinking appears as collapsed rows and approval requests (writing outside the workspace / running commands) can be answered in place: whole-plugin write route POST /api/code-server/ask/approve, constrained to pending in-process requests and the outcomes allowed-once/rejected; dialog-first for 5 minutes, handed back to the official card the moment the dialog closes, never auto-approved. There is no build step anywhere: lib/client.js (client half including the dialog panel) and the extension are committed hand-written source, devDependencies holds a single test-only entry, and the published package carries no frontend bundle. Also: the tab claims DSH file addresses (dsh-resource://file/**) by file type (setting claimExtensions), so produced-file chips, delivered-file previews and inline prose mentions open in the workbench; the IDE is a resident surface moved with Element.moveBefore instead of being remounted, so switching sidebar tabs does not reload it; opening the tab switches the right sidebar to fullscreen by default (setting fullscreenOnOpen); a workspace switch re-navigates the workbench with the new ?folder= without restarting the IDE process. Requires a DSH with the right-sidebar services (sidebarRightTabs/sidebarRight, >= 0.1.5-alpha.1); older DSH versions only get an upgrade notice on the settings page. Two serving modes: loopback port (default) or same-origin mount on DSH's own webServer (/code-server, protected by ctx.connection.requestRejection). The diff the bridge opens always has the real pre-write text on the left, taken from result.value.before in tools/post-execute into a bounded snapshot cache and fetched by opaque key over the read-only GET /old, so files that are not open in the editor work too. No code-server Node layer, no argon2, no C++ toolchain.
为 DeepSeek Harness 添加个性化设置
Context compression tool (context_compact): the agent writes the replacement checkpoint itself and hands it to the host compaction engine, which skips the LLM summarizer call; automatic compaction stays on the official engine.
Find, show/hide and arrange plugin controls in DeepSeek Harness Desktop and Web, with layout backup and restore.
订阅会员 OAuth 登录(ChatGPT 等):配置中心「订阅服务」页登录/注销,自动发现订阅模型列表,按登录状态注册提供商,支持思考强度选择
Desktop notifications for DeepSeek Harness (dsh): run completion, model questions, and approval requests. Zero runtime dependencies — all peers provided by the dsh CLI.
S7R System 7 Reimagined workstation plugin for DeepSeek Harness
Deterministic fault injection, autonomous resilience experiments, and machine-readable evidence for the DeepSeek Harness tool pipeline
语织:中文优先、用户可控的 AI 写作风格记忆层 / Chinese-first style memory for AI
Off-transcript clarification Host plugin for DeepSeek Harness
Structural code MCP server and batch CLI for compact TypeScript/JavaScript reads and hash-bound AST edits.
DSH composer model seat with accessible per-provider collapsible groups.
Token-efficient catgirl persona runtime for DeepSeek Harness