bigclawd/dsh-security-guard
C3Security guard for DeepSeek Harness (dsh): static scan for malicious code, prompt injection and token waste, runtime interception, /scan, plugin_scan, web panel and allowlist | DeepSeek Harness 安全守卫插件
★ 1+ · bigclawd/dsh-security-guard source on GitHub · this plugin in the registry
bigclawd/dsh-security-guard is a DeepSeek Harness plugin rated C3 — powerful capability combined with sensitive behavior. It patches the dsh runtime, runs code at install time, gates tool execution, executes system commands.
What it can do
| Capability | Flag | Evidence |
|---|---|---|
| patches the dsh runtime | runtime_patch | ./cordis.patch.yml |
| runs code at install time | install_script | postinstall: node src/postinstall.js |
| gates tool execution | tool_gate | hook: tools/pre-execute |
| executes system commands | exec | ×4 src, e.g. src/index.ts:5, src/types.ts:86 |
| uses eval / new Function | eval | ×3 src, e.g. src/types.ts:85, src/types.ts:85 |
| decodes base64 payloads | base64_decode | ×1 src, e.g. src/static/content.ts:124 |
| reads credential-class env vars | token_env | API_KEY |
Services it injects
invariants
Hooks it attaches
agent/pre-step fs/observed internal/service session/event tools/post-execute tools/pre-execute
Outbound domains
collect.evil.example.net
Environment variables it reads
API_KEY
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.