WSL043/dsh-codex-subscription
C3Use your ChatGPT Plus / Pro (Codex) subscription in DeepSeek Harness (DSH): GPT-6 & Codex models, images, web search and quota via ChatGPT sign-in — no OpenAI API key. Beta: control DSH from the ChatG
★ 100+ · WSL043/dsh-codex-subscription source on GitHub · this plugin in the registry
WSL043/dsh-codex-subscription is a DeepSeek Harness plugin rated C3 — powerful capability combined with sensitive behavior. It patches the dsh runtime, executes system commands, uses eval in development tooling the package does not ship, decodes base64 payloads.
Installable plugin — declares a dsh.bundle manifest
What it can do
| Capability | Flag | Evidence |
|---|---|---|
| patches the dsh runtime | runtime_patch | ./cordis.patch.yml |
| executes system commands | exec | ×7 in authored code, e.g. src/external-url.js:1, src/oauth-network.js:1 |
| uses eval in development tooling the package does not ship | eval_tooling | ×22 in development tooling the package does not ship, e.g. scripts/ui-smoke.mjs:67, scripts/ui-smoke.mjs:69 |
| decodes base64 payloads | base64_decode | ×5 in authored code, e.g. src/codex-images.js:122, src/original-image-download.js:13 |
| starts a network server in development tooling the package does not ship | net_server_tooling | ×3 in development tooling the package does not ship, e.g. scripts/phone-acceptance/lib.mjs:42, scripts/support/dsh-host.mjs:15 |
| reads credential-class env vars in development tooling the package does not ship | token_env_tooling | in development tooling the package does not ship: CODEX_AUTH_PATH, GH_TOKEN |
Tool registrations
24 confirmed registration point(s); analysis is incomplete.
lib/index.js:936tool factorylib/index.js:5105tool factorylib/index.js:5304tool factorylib/index.js:15055tool factorylib/index.js:15084tool factorylib/index.js:15108tool factorylib/index.js:16875tool factorylib/index.js:17935tools.registerlib/index.js:17959tools.registerlib/index.js:18360tools.registerlib/index.js:18386tools.registerlib/index.js:18471tools.registersrc/codex-images.js:365tool factorysrc/connectors.js:179tool factorysrc/connectors.js:197tool factorysrc/connectors.js:211tool factorysrc/index.js:422tools.registersrc/index.js:447tools.registersrc/index.js:795tools.registersrc/index.js:812tools.registersrc/index.js:859tools.registersrc/sketch-agent-tool.js:9tool factorysrc/subscription-preferences-tool.js:9tool factorysrc/voice-tools.js:8tool factory
1 source file(s) omitted by size/count/encoding limits; tool count is incomplete
Services it injects
attachments credentials llm loader settings tools web
Hooks it attaches
agent/assistant-stream agent/disposed agent/request agent/request-error approval/request loader/volatile-update session/created session/disposed session/event user-questions/request
Outbound domains
chatgpt.com api.openai.com auth.openai.com
Environment variables it reads
DSH_CODEX_REMOTE_DEBUG SystemRoot DSH_WEB_FETCH_PROVIDER NODE_TEST_CONTEXT DSH_CODEX_REMOTE_BASE
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.