NeoHotaru/dsh-knowledge-canvas

C3

Tree-shaped knowledge bases for DeepSeek Harness: a right-side canvas you edit by hand, plus tools the model reads and writes. Nodes carry a trigger line saying when they apply; a base can be injected

★ 10+ · NeoHotaru/dsh-knowledge-canvas source on GitHub · this plugin in the registry

NeoHotaru/dsh-knowledge-canvas is a DeepSeek Harness plugin rated C3 — powerful capability combined with sensitive behavior. It patches the dsh runtime, can rewrite the system prompt, executes system commands, uses eval / new Function.

Installable plugin — declares a dsh.bundle manifest

What it can do

CapabilityFlagEvidence
patches the dsh runtimeruntime_patch./cordis.patch.yml
can rewrite the system promptprompt_surfacesystemPrompt, webServer
executes system commandsexec×6 in authored code, e.g. test-bundle-install.mjs:6, test-bundle-install.mjs:6
uses eval / new Functioneval×4 in authored code, e.g. check-client-loads.mjs:694, test-cold-session-routes.mjs:63
decodes base64 payloadsbase64_decode×1 in authored code, e.g. client.js:173

Tool registrations

0 confirmed registration point(s); analysis is incomplete.

Services it injects

remote sessions sidebarRightTabs slots systemPrompt tools webServer

Hooks it attaches

agent/created agent/pre-step session/created tools/post-execute

Environment variables it reads

DSH_HOME

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.