DSH Marketplace

dsh-crew

ZSeven-W/dsh-crew

Dispatch work to DSH agents from Claude Code or Codex: native subagent progress, in-host worker sessions with per-tier presets, and a multimodal bridge for vision and image generation.

18516JavaScriptMITSource

Install

Add dsh-crew to DeepSeek Harness

Install into

via npm

Resolves a published tarball rather than cloning the repository, and installs without any extra setup. Swap `web` for your profile name if you run another one.

via GitHub · npm package

Installing from GitHub runs the project's build script, which pnpm blocks until you allowlist it — run the command once and pnpm prints the exact key to add under `allowBuilds` in ~/.dsh/profiles/web/pnpm-workspace.yaml.

What happened when we ran it

Installed cleanly when we ran it

Every command here is run in a throwaway container against a clean profile, and the result is whatever the harness recorded — not a guess from the source. Last run yesterdayon DeepSeek Harness 0.2.2-rc.5.

Show it in your README

install verified — dshmarketplace

For maintainers: the badge serves this listing's latest sandbox verdict, so a re-run updates it on its own — and it links readers to the full result here.

Due diligence

Before you install dsh-crew

  • Source of record: ZSeven-W/dsh-crew — present in the community registry that DSH's own plugin market installs from.
  • Licensed under MIT.

Security

What installing dsh-crew runs

Build scripts

No build script was blocked in our run.

Source: our sandbox install on DeepSeek Harness 0.2.2-rc.5, yesterday.

Install hooks

package.json declares no preinstall, install or postinstall script.

Source: package.json on GitHub, read at our last sync. prepare is not recorded; it runs only on a GitHub install, and shows up under build scripts when pnpm blocks it.

README mentions

shell commands (bash, exec, spawn …)

Source: a keyword match on the README — a hint about what the plugin touches, not an analysis of its code.

Not an audit. These are the record of one install and two text scans; nobody has read this plugin's code on your behalf. Plugins run with your agent's permissions.

The AI take

What it is — DSH Crew plugin lets DeepSeek Harness agents dispatch work from Claude Code or Codex while keeping native subagent progress UI.

Who it is for — When handling multi-step AI agent tasks in Claude Code, DSH Crew plugin fits dispatching to DSH agents as subagents. Users of Grok do not need it, as they have no need for external tool integration.

Watch out — Sandbox tested through: successfully installed and registered in a fresh profile. Executes shell commands. No obvious issues found.

The verdict — I would install it because it adds native progress UI and vision bridging, but only if the workflow involves external tool dispatching.

Generated by grok-4.6, and a starting point rather than a verdict. Where it says a plugin installs or does not, that is from a real run in a clean profile — everything else is read off the repository. Trust the source over this.

What dsh-crew does

Abstract diagram of worker dispatch branching through tier gates into sessions, image channels and terminal commands.

dsh-crew is a DeepSeek Harness plugin that dispatches work from Claude Code, Codex, Antigravity or Grok to DeepSeek Harness agents while preserving the host's native subagent progress view.

The host invokes the MCP command dsh_run_worker with a tier, effort, cwd and optional worker. flash and pro select the configured agent capability; tier_policy can constrain dispatches, while escalate_on_failure can retry a failed flash job on pro. When the DeepSeek Harness hub is reachable, each job runs as an in-host session and appears in the Web UI. Otherwise, dsh-crew uses the standalone dsh-jsonrpc-agent runtime and worker.cordis.yml. Progress events, tool calls and token usage are exposed to the host and listed on the jobs board.

The plugin is intended for developers who want external coding agents to delegate to DeepSeek Harness workers, including multimodal calls through describe_image and generate_image. It is a poor fit for workflows that only need DeepSeek Harness's existing sessions, or where external CLI execution is not acceptable. The optional agy and grok worker paths invoke those CLIs, and custom providers can run a local command template. This gives dsh-crew a terminal surface: it can reach the host terminal and configured commands with the agent's permissions. Worker nesting and workspace contention are refused rather than queued.

dsh-crew documentation

Configuring dsh-crew

DeepSeek Harness creates workers from its configured agent roster when tier is flash or pro. The dispatch controls documented by the README are:

Key Values or inputs Behaviour
tier_policy a tier policy Clamps every dispatch at the tool layer
escalate_on_failure enabled or disabled Retries a failed flash run once on pro
effort off through max Sets the worker's reasoning effort
cwd a working-directory path Groups an in-host session and applies workspace checks
worker agy or grok Explicitly selects the corresponding external CLI

The README does not specify defaults for these settings. DeepSeek Harness sessions use the Agent preset selected for each tier.

dsh-crew Commands

The MCP command dsh_run_worker accepts tier, effort, cwd and worker. A normal dispatch runs inside the DeepSeek Harness hub when that hub is reachable. The fallback is the standalone dsh-jsonrpc-agent runtime, using worker.cordis.yml. The agy and grok values make the selected external CLI run the task instead of a DeepSeek Harness worker.

How dsh-crew behaves

DeepSeek Harness exposes each in-host worker as a first-class session in the Web UI, grouped by working directory. The host integration reports dispatch count, current step, tool calls and token usage. The jobs board retains running and finished jobs, shows tier, effort, progress and tokens, and marks a job as an orphan ghost if it disappears during a hub restart.

Before spawning, dsh-crew rejects worker-to-worker chains beyond origin depth 3, cycles, and a second job targeting a workspace already held by another job. describe_image prefers deepseek-v4-flash-vision-exp when a key is available, then can use configured CLIs or an OpenAI-compatible provider. generate_image uses the same provider routes.

dsh-crew Requirements and limits

DeepSeek Harness was tested with 0.1.1-rc.1; the plugin release shown in the README is 0.1.0-rc.11. Vision and image generation depend on an available key, CLI or configured provider. DSH models are text-only, so multimodal work uses the bridge rather than the worker model directly. The plugin can also run configured local command templates, which is the relevant terminal permission surface.

Written from the project's own documentation and kept in sync with it. Where the two disagree, the source is authoritative — read the README on GitHub

Same category

Alternatives to dsh-crew

dsh-loopx-plugin

huangruiteng

6.7k

LoopX, a provider-neutral, local-first state kernel and control plane for long-horizon agents: keeps Goal, Todo, gate, evidence, quota, recovery, and handoff state above DeepSeek Harness, while the plugin bootstraps the CLI and skills, admits bounded same-session continuation, and adds a loopback GoalBar for the exact bound loop.

No one-line install — this plugin lives inside a larger repository and publishes no npm package.

GitHub sourcePythonyesterday

AI review

dsh-loopx-plugin

What it is — DeepSeek Harness plugin that bootstraps LoopX CLI and integrates skills.

Who it is for — For managing long-horizon agent tasks using DeepSeek Harness, it adds state management. Users running agent tasks on non-DeepSeek Harness frameworks won't need this plugin.

Watch out — No obvious issues identified. Static checks show it will execute shell commands. Sandbox testing has not been conducted.

The verdict — I would install it because it enables long-horizon state management on DeepSeek Harness, but only for tasks requiring bounded loop continuation.

Generated by grok-4.6, and a starting point rather than a verdict. Where it says a plugin installs or does not, that is from a real run in a clean profile — everything else is read off the repository. Trust the source over this.Read the source
1Source
6.4k

Config-only bundle that mounts Ouroboros through the DSH MCP client, exposing 36 interview, Seed, execution, evaluation, and evolution workflow tools in DSH.

No one-line install — this plugin lives inside a larger repository and publishes no npm package.

GitHub sourcePython3d ago

AI review

dsh-plugin

What it is — This DSH plugin mounts Ouroboros through the DSH MCP client, exposing 36 interview, Seed, execution, evaluation, and evolution workflow tools in DeepSeek Harness.

Who it is for — Users performing workflow tasks in DeepSeek Harness benefit from this plugin. People who do not need to mount external Agent OS via DSH MCP do not need this plugin.

Watch out — It is a subdirectory plugin with no single install command. It will execute shell commands. No sandbox testing was performed; no obvious pitfalls found.

The verdict — I would install it because it integrates Ouroboros directly for DeepSeek Harness.

Generated by grok-4.6, and a starting point rather than a verdict. Where it says a plugin installs or does not, that is from a real run in a clean profile — everything else is read off the repository. Trust the source over this.Read the source
1Source

dsh-loopx-plugin

loopx-project

6.3k

LoopX, a provider-neutral, local-first state kernel and control plane for long-horizon agents: keeps Goal, Todo, gate, evidence, quota, recovery, and handoff state above DeepSeek Harness, while the plugin bootstraps the CLI and skills, admits bounded same-session continuation, and adds a loopback GoalBar for the exact bound loop.

No one-line install — this plugin lives inside a larger repository and publishes no npm package.

GitHub sourcePython1mo ago

AI review

dsh-loopx-plugin

What it is — The dsh-loopx-plugin provides a LoopX control plane for DeepSeek Harness, managing durable goals, todos, evidence, and handoffs for long-horizon agents.

Who it is for — When building coding or research agents with DeepSeek Harness that require cross-session Goal and Todo state, this plugin can be installed. For users who only need single-session instant task completion, the durable state management provided by this plugin falls outside their requirements.

Watch out — No obvious issues found. From source installation, the build script requires manual approval to execute.

The verdict — I would install it because it maintains a bounded loop GoalBar for DeepSeek Harness agents, which is necessary for long-horizon work.

Generated by grok-4.6, and a starting point rather than a verdict. Where it says a plugin installs or does not, that is from a real run in a clean profile — everything else is read off the repository. Trust the source over this.Read the source
1Source
3.5k

Embeds the active installed Codex Taskboard runtime in the DeepSeek Harness sidebar, using its launcher runtime descriptor instead of a fixed port.

No one-line install — this plugin lives inside a larger repository and publishes no npm package.

GitHub sourceJavaScript7d ago

AI review

deepseek-harness

What it is — Integrates Codex Taskboard runtime into DeepSeek Harness sidebar.

Who it is for — Users who embed Codex Taskboard into DeepSeek Harness sidebar are suitable to install this plugin. Users who do not want to provide an API key or token are not suitable to install it, because it requires providing one.

Watch out — No sandbox testing was performed. It requires providing an API key or token and may execute shell commands. No obvious issues found.

The verdict — I would not install it because there is no single command to install it.

Generated by grok-4.6, and a starting point rather than a verdict. Where it says a plugin installs or does not, that is from a real run in a clean profile — everything else is read off the repository. Trust the source over this.Read the source
2Source

Installs a dedicated AI novel-writing preset and workbench: revisioned local project assets, a compact side drawer, and native approval-gated single-file changes.

No one-line install — this plugin lives inside a larger repository and publishes no npm package.

GitHub sourceTypeScript3d ago

AI review

dsh-ai-novel-writer

What it is — Adds an AI novel-writing preset and workbench to DeepSeek Harness.

Who it is for — Users who are writing long novels in DeepSeek Harness and want AI assistance for chapter generation with manual review are suitable for installing it. Users who require automated batch generation or integration with mature editors should not, as it only supports a minimal manual review workflow.

Watch out — No obvious issues found. This is an early MVP version with capabilities only about 10% of the desktop version. It only supports the minimal creation chain with manual approval.

The verdict — I would not install it, because it lacks the full project tree and batch workflows of the desktop version.

Generated by grok-4.6, and a starting point rather than a verdict. Where it says a plugin installs or does not, that is from a real run in a clean profile — everything else is read off the repository. Trust the source over this.Read the source
Source

dsh-agent-teams

NanmiCoder

2.2k

AgentTeams multi-agent teams.

Installed cleanly when we ran it

npm packageJavaScript1mo ago

AI review

dsh-agent-teams

What it is — Turns the current DeepSeek Harness session into a captain runtime: continuable sub-agents, a dependency-aware task state machine, direct member mailboxes, and a live Web activity panel.

Who it is for — You are in DeepSeek Harness splitting one goal into dependency-ordered tasks for continuable sub-agents, and you want the Web panel for the task DAG and member progress. Do not install if that same session must keep writing the implementation: the plugin parks it in the captain seat—create the team, assign, consolidate.

Watch out — The recommended host is still DeepSeek Harness 0.2.1-rc.2, an RC; 0.1.25 exists to stop that host from rejecting the install. The plugin executes shell commands. Opt-in quality gates audit at completion and do not intercept host writes. A fresh-profile sandbox install registered with the harness; the desktop app ships its own core, so upgrading the global CLI does not move it.

The verdict — Worth it only if staging a roster and DAG first is acceptable—scheduling does not start until Approve & Run.

Generated by grok-4.6, and a starting point rather than a verdict. Where it says a plugin installs or does not, that is from a real run in a clean profile — everything else is read off the repository. Trust the source over this.Read the source
1Details