Skill Trust Decision

job-agent

远程脚本执行风险(curl|bash)未在SKILL.md中明确警告,且无版本锁定验证机制。功能本身是合法的求职辅助CLI,但安装方式存在供应链风险。

Install decision first Source: ClawHub Scanned: Sep 23, 2026
Files 2
Artifacts 7
Violations 0
Findings 3
Most direct threat evidence
01
用户执行远程安装命令 Entry · SKILL.md
02
攻击者劫持main分支,注入恶意代码到install.sh Escalation · SKILL.md
03
恶意安装脚本执行,可能安装后门或窃取凭证 Impact · SKILL.md

Why this conclusion was reached

2/4 dimensions flagged
Pass
Declared vs actual capability

Declared resources and inferred behavior are broadly aligned.

Block
Hidden execution and egress

1 high-risk artifacts or egress signals were extracted.

Block
Attack chain and severe findings

The report includes 3 attack-chain steps and 1 severe findings.

Review
Dependencies and supply chain hygiene

Dependency information is incomplete, so supply-chain confidence stays limited.

Attack Chain

01
用户执行远程安装命令

Entry · SKILL.md:216

02
攻击者劫持main分支,注入恶意代码到install.sh

Escalation · SKILL.md:216

03
恶意安装脚本执行,可能安装后门或窃取凭证

Impact · SKILL.md:216

What drove the risk score up

远程脚本执行未声明风险 +20

SKILL.md第216行curl|bash管道执行远程脚本,但文档安全合约部分未明确说明安装风险和验证步骤

无安装脚本版本锁定 +15

使用main分支URL,无commit hash或tag锁定,存在供应链劫持风险

PowerShell安装同样未说明风险 +10

Windows安装命令irm|iex同样为远程脚本执行模式

Most important evidence

High Supply Chain

远程脚本执行无版本锁定

安装命令使用main分支URL `https://raw.githubusercontent.com/jiyangnan/AgentMesh-JobAgent/main/scripts/install.sh`,无commit hash或tag锁定。攻击者可劫持main分支植入恶意代码。

SKILL.md:216
使用带commit hash的固定URL,或要求publisher在文档中说明如何验证安装脚本完整性
Medium Doc Mismatch

安全合约未涵盖安装风险

文档包含详细的安全合约(用户授权、凭证处理、投递确认),但未对远程安装命令的风险进行说明和缓解指导。

SKILL.md:49
在安全合约中增加安装验证说明,如:'安装前请审查脚本内容或使用版本锁定URL'
Low Supply Chain

Windows安装命令同样存在风险

PowerShell安装 `irm https://.../install.ps1 | iex` 与macOS/Linux命令具有相同的供应链风险

SKILL.md:219
同上,使用版本锁定的URL

Declared capability vs actual capability

Shell Pass
Declared WRITE
→
Inferred WRITE
SKILL.md:216-220 包含安装命令
Network Pass
Declared READ
→
Inferred READ
多个外部URL用于配置和工作流程
Browser Pass
Declared NONE
→
Inferred WRITE
文档描述使用CLI驱动浏览器进行平台操作
Filesystem Pass
Declared READ
→
Inferred READ
文档声明读取简历文件~/.jobagent状态

Suspicious artifacts and egress

Critical Dangerous Command
curl -fsSL https://raw.githubusercontent.com/jiyangnan/AgentMesh-JobAgent/main/scripts/install.sh | bash

SKILL.md:216

Medium External URL
https://jobagent.agentmesh360.com/

SKILL.md:8

Medium External URL
https://agentmesh360.com/app/

SKILL.md:173

Medium External URL
https://agentmesh360.com/workbench/

SKILL.md:176

Medium External URL
https://agentmesh360.com/app/?lang=zh-CN#pricing

SKILL.md:237

Medium External URL
https://clawhub.ai/user/jiyangnan

skill-card.md:9

Medium External URL
https://clawhub.ai/jiyangnan/skills/job-agent

skill-card.md:40

Dependencies and supply chain

There are no structured dependency warnings.

File composition

2 files · 379 lines
Markdown 2 files · 379 lines
Files of concern · 2
SKILL.md Markdown · 319 lines
远程脚本执行无版本锁定 · 安全合约未涵盖安装风险 · Windows安装命令同样存在风险 · curl -fsSL https://raw.githubusercontent.com/jiyangnan/AgentMesh-JobAgent/main/scripts/install.sh | bash · https://jobagent.agentmesh360.com/ · https://agentmesh360.com/app/ · https://agentmesh360.com/workbench/ · https://agentmesh360.com/app/?lang=zh-CN#pricing
skill-card.md Markdown · 60 lines
https://clawhub.ai/user/jiyangnan · https://clawhub.ai/jiyangnan/skills/job-agent

Security positives

文档声明的功能与实际行为一致,无阴影功能
凭证处理明确要求用户确认和本地配置,无外传行为
投递操作需要预览、授权ID和最终确认,用户控制充分
skill-card.md已知风险并提供缓解指导
CLI工具本身是开源的(AgentMesh-JobAgent)