Plugins

Ferramentas de estabilidade de testes: execute repetidamente suítes vitest/jest/pytest/node:test, classifique casos instáveis, mantenha histórico, um manifesto de quarentena e autoverificação de saúde.

Guarda de consistência entre documentação e código: audita cabeçalhos de versão, tabelas de changelog, árvores de diretório, contagens de módulos e testes e referências cruzadas entre documentos em busca de desvios, com sugestões de correção somente leitura ordenadas por severidade.

Entrega trabalhos pontuais à CLI do Cursor logada na mesma máquina, com diagnóstico de prontidão da CLI e de proxy, configurações de modelo filho e um teste de conectividade somente leitura.

Verificador de atualizações (edição npm): uma página de configurações que compara as versões local e mais recente do @deepseek-ai/dsh e de cada plugin instalado, com lembrete de ponto vermelho na navegação e comandos de atualização prontos.

Arnês de regressão de contrato para autores de plugins: um adaptador LLM roteirizado conduz o pipeline real em processos filhos isolados com asserções de eventos de sessão (camadas load/contract/behavior/fault); sem chave de API, pronto para CI.

Guarda de inicialização que repara logs de sessão corrompidos, pré-inspeciona a composição de plugins, verifica por vm os artefatos de cliente, faz smoke test do apply() do host e põe em quarentena bundles que causam travamentos, para que um plugin quebrado nunca inutilize a inicialização.

Empacota uma biblioteca de skills de metodologia de engenharia como um único plugin Cordis: 14 Agent Skills (planejamento, testes, depuração, revisão, entrega) registradas em ctx.skills para o DeepSeek Harness.

Registra a skill Inspiration Deck Workshop para o DSH: apresentações HTML estáticas locais (6 modelos, mais de 25 layouts, 23 temas e showroom de movimento) com CLI de validação e exportação para PNG/PDF e testes de fumaça, sem dependências de runtime.

Cria, modifica e injeta mensagens de assistente de teste e importa/repara logs de sessão session.jsonl(.zstd) em novas sessões.

Dobra o raciocínio, as chamadas de ferramentas e o contexto de cada turno do DSH Web em caixas de processo com bordas, cortadas a cada trecho de texto do modelo; uma caixa recolhida mostra seus dois itens mais recentes, as linhas empilhadas acima de sua borda abrem aquela caixa, a própria borda a recolhe de volta, e o diálogo do usuário, as perguntas interativas e o texto do modelo ficam fora.

Ferramenta estruturada de execução de testes (test_run): detecta automaticamente Vitest/Jest/pytest/node:test, executa os testes e resume as falhas para o modelo

A CLI complementar que faltava para o DeepSeek Harness (dsh): gerencie servidores MCP com um único comando — add/list/remove/test com verificações de conexão em modo dry-run, além de importação única a partir do Claude Code / Codex. Também se monta como plugin do dsh, para que o próprio agente possa gerenciar o MCP

Profiler de contexto de requisições para o DeepSeek Harness — veja o que mudou entre as requisições ao modelo e como o reaproveitamento de cache mudou junto

DeepSee: visão para o DeepSeek Harness, roteamento multimodelo e autoverificações visuais com Gemini antes da entrega

DSH 测试自动修复：给 agent 一个 run_tests 工具，跑测试并把结构化失败（文件/行号/断言）直接交给它修复。自动识别 pytest / jest / vitest / cargo / go / npm，Windows 中文断言不乱码，同一失败反复出现时自动停止。

DSH plugin: inventory your Windows dev/test toolchain, spot the gaps, and fill them with one click.

DSH plugin: deterministic, offline Python/Django production-report diagnostics — traceback root cause, pytest summaries, unapplied migrations, N+1 SQL, manage.py check --deploy and Celery backlog. Does not execute Python, start subprocesses or touch the n

Pre-claim gate: a commit, push or tag claim like "tests pass" is checked against the session evidence ledger before it is written.

Outbound route pool for DeepSeek Harness. Ships 15 seed routes plus a 20-entry candidate pool covering 7 target kinds (GitHub raw/codeload/main, npm registry, PyPI index, HuggingFace, general). Routes are verified before use: upstream IP lists go through a TLS handshake with full certificate-chain validation (a bare TCP check can be answered locally and report dead IPs as healthy), and failures advance a cooldown ladder of 2/5/15 minutes with the fastest healthy route picked per target kind. Routes are applied by rewriting the URL with one of four rules (prefix, registry, index, host-replace), reaching git, pip, npm, pnpm, gh, curl and wget through three paths: the shell seam (download commands the model runs through the bash tool are intercepted and rewritten automatically; commands wrapped by the pwsh tool, and fetches made by the host itself or the renderer, are not covered), two registered tools the model can call directly (line_pool_status, line_pool_run), and child-process spawn. Multi-line commands keep their line breaks and only the download segment is touched. URLs carrying credentials (user:token@host) are never rewritten, so tokens are not forwarded to third-party mirrors. No global config is modified. A sidebar panel lists every route with latency, failure reason and cooldown countdown, supports per-route retest and enable/disable, and shows live TUN/proxy status from network-adapter and loopback port checks. The candidate pool is re-verified every 24 hours so newly usable routes are added and dead ones dropped. 570 tests, 87% line coverage.

Keep a DSH profile identical across machines through any WebDAV server. The plugin collects the five parts that make up a setup — models (the llm-pi-ai line of cordis.patch.yml), defaults (the agent-default-model line), patch (the remaining loader lines), plugins (package.json dependencies plus dsh.profile.bundles) and dirs (extra absolute-path directories picked in the panel, carried whole and restored to the same path, add-only, at most 5 with 2000 files / 8 MiB each) — encrypts them on this machine with a passphrase (AES-256-GCM, PBKDF2-SHA256 at 600000 iterations) and uploads one ciphertext file to <address>/dsh-webdav/dsh-profile.prop, so the server only ever holds the sealed copy (the older dsh-profile.dshsync name is still read). Merging is per part and three-way against the last synced baseline; a conflict keeps the newer copy and archives the losing one under sync/ (latest 5 per tag), every overwrite is sealed beforehand, and deletions propagate except in the plugins part, on the first sync, and in the dirs part, where dropping a directory on one machine only stops syncing it there and never deletes its files. Uploads carry ETag If-Match and If-None-Match so a concurrent change is refused instead of being clobbered. Sync runs on demand or on a 3/15/30/60 minute timer from a settings panel, and an optional local forwarder (off by default, loopback-only) re-serves the same target with locally generated credentials, rewritten PROPFIND hrefs and .prop aliases for extensionless resources. The local state lives in <profileDir>/.dsh-webdav-sync/ (sync.json is written 0600 and holds the WebDAV password and the passphrase).

Plugin development toolset: mark files read by shell commands, sync code into a profile, mock a plugin backend, preview a plugin frontend, verify plugin loading, spin up an isolated test instance, run named shell shortcuts, and list the plugin's tools.

Turns what a design, plan or document claims into a model and decides it: UML diagrams rendered, read and reviewed; state machines, protocols and data models checked as models; migrations tested for coverage and rollback symmetry; concurrency claims mined and routed to the checker that can decide them; dependency graphs audited; reports diffed.

Generic monitor plugin for DSH Desktop: an HTTP polling rule engine. Poll any endpoint on schedule (multi-step, variable extraction, JS condition expressions); when conditions hit, wake a visible agent session to act — zero model calls otherwise. Configure everything in a sidebar GUI panel: create/edit/pause/test/fire-now/run history, same UX as the built-in scheduler plugin.

Test work framework for DSH that scaffolds a product, module and version directory tree, registers seven test_framework_* tools and four runtime skills, and adds a prompt-level gate that reports which requirement-document splits, versions, related documents and environment credentials are still unrecorded before testing starts. Requires the playwright MCP server (no fallback).