中文

← Back to list

dsh-prompt-antivirus

DeepSeek Harness Other Low risk

Global prompt-injection / context-virus defense for DeepSeek Harness: scans tool arguments, tool results, pre-model messages and the outbound stream; quarantine/block/monitor modes, canary trap, and an evolvable on-disk signature library with learn/import/export.

全局提示注入 / 上下文病毒防御：扫描工具参数、工具结果、进模型前消息与出站流；隔离/阻断/监控三模式、金丝雀陷阱、可演进磁盘签名库（学习/导入/导出）。

How to install

DeepSeek Harness dsh plugin add github:QinpanWan/dsh-prompt-antivirus

About

dsh-prompt-antivirus — 全局防上下文病毒感染 运行时防御：扫描 · 隔离 · 金丝雀陷阱 · 人工门控 · 审计，覆盖 dsh 全部预设与子代理。 原理移植自 MIT 项目 openclaw-prompt-antivirus（github.com/QinpanWan/openclaw-prompt-antivirus），按 dsh（cordis）钩子面接线。 安装 sh dsh plugin --profile web add github:QinpanWan/dsh-prompt-antivirus 装完重启 dsh web 生效（日志出现 [prompt-antivirus] 已加载）。插件挂载在 profile 层， 对所有预设与子代理全局生效。源码安装 / 手动部署见文末「接入」。 问题 提示注入是 LLM 智能体的核心弱点：模型分不清「指令」与「数据」——藏在网页检索结果、文件内容、 [CRON TASK] / [SCHEDULE REMINDER] 到期文本里的恶意文字，可以悄悄改写行为、外传数据或执行破坏性操作 （即「上下文病毒」感染）。本插件在数据进入模型前后各加一道防线。 拦截点（dsh 钩子 ↔ OpenClaw 对照） | 防线 | dsh 钩子 | 行为 | | --- | --- | --- | | 工具参数扫描（直接注入） | to…

GitHub ↗

Recommendation signals

44 Tool quality · Based on stars, downloads, maintenance, security and docs
– User interest · Adjusted by in-site views, install copies and download clicks
44 Overall
0views
0unique visitors
0install copies
0download clicks
0outbound clicks

Meta

License
MIT
Language
JavaScript
GitHub stars
1
mo. downloads
–
Last push
2026-08-31
Created
2026-08-31

Links

GitHub ↗ Report issue ↗

Basic safety check

Findings
curated 收录但无 npm 包/安装命令
Sources
curated:awesome-dsh-plugin.com, curated:awesome-dsh-plugin/awesome-dsh-plugin
Topics
dsh, dsh-plugin, harmonyos, prompt-injection, security

Related plugins

DeepSeek Harness Featured
Score74

dsh-secure-audit

PensiveFei/dsh-secure-audit

Read-only security and compliance plugin for DeepSeek Harness: prompt-injection detection, Chinese-PII redaction, and a local configuration audit with redacted, reproducible reports.

☆ 87 ↓ 1.7K Other ↗
DeepSeek Harness
Score69

dsh-auto-approve

Jiao-XXX/dsh-auto-approve

Adds an `auto` permission preset between workspace-write and danger-full-access: a classifier grants routine sandbox escalations once, while dangerous or uncertain requests still go to a human.

☆ 14 ↓ 1.8K Other ↗
DeepSeek Harness Featured
Score66

dsh-plugin-vet

wulun811/dsh-plugin-vet

Plugin trust pipeline for DeepSeek Harness: deterministic static scan with verdicts, opt-in runtime guard with honeypot lures, agent audit-protocol skill, and a browser shield status light. Alarm-only, never an enforcer.

☆ 3 ↓ 4.6K Other ↗
DeepSeek Harness Featured
Score65

dsh-agent-approval

MoonlitDropOfBlood/dsh-agent-approval

An independent approval subagent judges every sandbox escalation, with a configurable model and an audit log.

☆ 3 ↓ 1.7K Other ↗