Plugins

All-in-one combo patch for STARDUSTLC's 18 DSH plugins: one command injects shared default configs across office, media, DevOps and notification presets; each component installs directly via a one-line add command (flat design, no git subdependencies).

Intercepts turn-stopping to force one critique-and-improve round after each completed turn before the agent stops.

Deterministic Reasoning Engine (DRE) for dsh: dre__* tools for knowledge verification (three-stage discrimination), deterministic cognition loops, constraint solving, mental models and synapse memory - strengthens information certainty.

Keeps DSH agents from forgetting your requirements during long tasks. It saves important conditions and completion evidence locally, brings them back after context compaction or session resume, and stops partial work from being reported as the whole task done.

Role-based plugin recommender for DSH: asks the user's role (developer, designer, writer, researcher, ops, student...), recommends suitable plugins, and audits each with a local static security scan (dangerous code patterns + 0-100 risk score) and a reputation check (npm downloads / GitHub stars).

Skin author workbench: official token browser (169 --dsw-* variables with Chinese descriptions and module-usage tags), live try-on, durable skin save, theme freeze, seven spectrum presets with SVG backdrops and custom backdrop upload.

Aemeath theme for the dsh web GUI — sakura pink × icy cyan × pearl white × deep navy, dual light/dark, animated pulsing composer border, and wallpaper backdrops

Unified appearance coordinator: theme color presets, UI fonts, UI font size, plus per-workspace markdown font size / line-height / background, with a coordination layer that stops UI scaling from leaking into chat body text.

求职指挥中心：JD 分析评估（A-F 评分）、简历优化、求职管线与面试准备。受 santifer/career-ops（65k★ MIT）启发的中文精简原创技能。

DSH web plugin: desktop notification, chime, and title flash when a session stops — finished output or waiting for human input. Click a notification to jump to that session.

Spend gate for DeepSeek Harness: refuses over-budget model calls on the llm/stream waterfall, before they are billed. Auto-downgrade past a soft limit.

Adversarial code review — parallel finders inspect a diff through separate lenses (correctness, lifecycle, contract, security), then each finding goes to independent verifiers tasked with refuting it, and one refutation drops it.

Paddock: constrained local filesystem MCP server for agents — text read/write, directory ops, filename and content search confined to configurable zones, with zero runtime dependencies and a dsh bundle.

Compaction backend replacing LLM summarization with a deterministic semantic extractor (keeps code/paths/commands, drops chatter) plus 28.4x KV-compression accounting.

FinOps de agentes / análise de tokens para o DeepSeek Harness: coleta de uso, motor de custo sensível ao tempo, análise de cache e detalhamento por sessão

Proteção contra desvio de premissas pós-compactação: injeta um aviso único quando um resumo de compactação descarta uma âncora literal crítica