Plugins

Security-audit facts AI coding agents cannot compute, plus a verdict gate: euthyna history attributes deleted lines to commits and flags those from security fixes (--origins finds the first introducer), euthyna coverage reports which changed symbols no test ever invoked, and euthyna gate checks each finding against six gates, downgrading any without evidence to an observation.

Biomedical research plugin for DeepSeek Harness (dsh): biomcp MCP server tools, scientific plotting skills, and deep-research subagent

Canonical project memory for DeepSeek Harness: one governing article per asset, an append-only journal beneath it.

Community skill marketplace for DeepSeek Harness: browse, search, and one-click install skills from awesome-skills-cn and the wider Claude Code / OpenClaw skill ecosystem.

Local-first HTML review and annotation workspace for coding agents: opens an isolated copy of a file in a browser editor, collects text edits, format changes and element/area comments, and returns a durable handoff the agent can claim and resolve.

“What should we build next?” There’s a skill for that.

CodeAct agent loop for the DeepSeek Harness: a persistent IPython session as the model's action space, with harness tools bridged in as a virtual `__dsh__.tools` module.

Agent-agnostic skill that turns task input files into reusable structural briefs and a per-task searchable index (OpenAI Codex, Claude Code, DeepSeek Harness)

Compare DeepSeek Harness, Codex, and Claude Code instruction surfaces for the same workspace.

Branch-role guard for AI coding agents: configurable integration/preview/production/archive roles, blocking direct push, force-push and delete on protected branches and agent merges into production/archive.

DSH plugin: auto-import local AI CLI conversations (Claude Code, Codex CLI, Cursor Agent) as real DSH sessions

Run the main session model on a local Claude Code subscription through the official Claude Agent SDK, taking over the llm/stream route for the claude-code provider, with token-level streaming, session resume, and DSH tools bridged in over MCP so they keep DSH sandbox and approval.

Scans MCP servers configured in Claude Code, Codex and CodeBuddy and probes real connectivity via stdio and streamable-http initialize handshakes, with enable/disable toggles and a /mcp control panel.

Subagent delegation provider that runs any Claude-compatible CLI binary pointed at by a configurable executablePath field.

Discovers Codex and Claude project directories and registers them as shared DeepSeek Harness workspaces.

Exposes the DSH dynamic-Cordis toolset to Claude Code over MCP (streamable HTTP on the existing web port, no new listener): ten tools to list, inspect, define, run, stop and permanently remove dynamic plugins in a live DSH session, with the same approval flow the GUI uses for packages that ship a client half. Bearer auth is mandatory — the token is configured on a DSH settings page, the endpoint answers 503 rather than serve without one, accepts loopback only and never emits CORS headers.

The task-center plugin set for DeepSeek Harness in one package: task ledger, model tools, human commands, wake patrol, quota-resume guard, idle-session mining, the reaper, scheduled sends, and the web kanban.

DeepSeek Harness (DSH) plugin that delegates coding tasks to the Claude Code and Codex agent CLIs and streams the whole run live — autonomous control, multi-account, automatic install, git worktrees, and any Anthropic-compatible endpoint.

Execute Claude Code e Codex como painéis de terminal ao vivo ao lado do dsh: uma visualização dividida na interface web, botões para posicioná-los e ferramentas que permitem ao agente dsh delegar trabalho a eles e ler suas telas.

Plugin out-of-tree para DeepSeek Harness: dá olhos a um modelo de codificação somente texto roteando imagens para uma rota Qwen-VL (DashScope) via ctx.llm e retornando texto.

Meta-skill de pareamento tarefa-skill com uma escada de preguiça para carregamento mínimo, uma tabela de roteamento editável pelo usuário e um fluxo de instalação com quarentena, varredura SkillSpector e aprovação humana.

Consultas headless ao Claude Code com papéis de advisor, reviewer e designer, retornadas como respostas de referência.

Traga seu histórico existente do Claude / ChatGPT para o DSH: importação instantânea sem tokens do memories.json dentro de uma exportação do Claude, transcrições locais do Claude Code sem necessidade de exportação e conversas destiladas pelo modelo que você já configurou no DSH. Cada memória carrega evidência literal verificada contra a fonte por código — uma paráfrase é descartada e contabilizada, e evidência encontrada apenas nos turnos do assistente é sempre encaminhada para revisão em vez de armazenada.

Pacote com 35 skills voltado a não desenvolvedores (bundle de profile do DSH, não um preset de Agent): orquestração + engenharia + produto + interface/lançamento. Basta `dsh plugin add coding-coach` para instalar as skills em qualquer profile; o Agent completo "Coding Coach" está no diretório dsh/coding-coach do repositório