ZSeven-W/dsh-crew

C3

DeepSeek Harness (DSH) plugin: dispatch work to DSH agents from Claude Code / Codex — native subagent progress, in-host worker sessions with per-tier presets, and a multimodal bridge that lends the te

★ 100+ · ZSeven-W/dsh-crew source on GitHub · this plugin in the registry

ZSeven-W/dsh-crew is a DeepSeek Harness plugin rated C3 — powerful capability combined with sensitive behavior. It patches the dsh runtime, executes system commands, eval in build output only, decodes base64 payloads.

Installable plugin — declares a dsh.bundle manifest

What it can do

CapabilityFlagEvidence
patches the dsh runtimeruntime_patch./cordis.patch.yml
executes system commandsexec×35 in authored code, e.g. scripts/smoke-bundle.mjs:16, scripts/smoke-bundle.mjs:68
eval in build output onlyeval_bundled×1 in build output only, e.g. lib/server.mjs:2231
decodes base64 payloadsbase64_decode×1 in authored code, e.g. src/multimodal.mjs:260
starts a network servernet_server×2 in authored code, e.g. scripts/verify-wpc9-guardrails.mjs:336, scripts/verify-wpc9-guardrails.mjs:383
reads credential-class env varstoken_envDEEPSEEK_API_KEY

Services it injects

agentDefaultModel agents attachments llm locale sessions slots tools

Hooks it attaches

agent/pre-step locale/change

Outbound domains

api.example.com api.deepseek.com

Environment variables it reads

DEEPSEEK_API_KEY CLAUDECODE CLAUDE_CODE_ENTRYPOINT DSHPLUGIN_CREW_HUB DSH_CREW_HUB DEEPSEEK_BASE_URL

How to read this

Levels measure capability surface and transparency, not maliciousness. A C3 plugin can be entirely legitimate — a desktop shell genuinely needs subprocesses. The point is that you can see this before installing. See the levels explained and how dsh plugins work.

Findings come from static analysis of shipped code; nothing is executed. Think a flag is wrong? Open an issue — every flag cites the file and line it came from.