Back to skill

Security audit

dev-mentor

Security checks for vulnerabilities and agentic risk

Overview

The skill is marked suspicious because its deployment guidance includes steps that could cause lockout or affect data without adequate safeguards; no clear malicious behavior was reported.

Consider the suspicious verdict before installing. The main reported concerns are missing safeguards around database migrations and SSH hardening, which could affect existing data or cut off server access. The scan was partial because some local references were unresolved, so it does not establish that all referenced material was reviewed.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (5)

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
85% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · tracks/deployment.md (reported line 30)May include surrounding context.

md
### 步骤 2.1 — Linux 基础

1. **常用命令**
   - 文件操作:ls / cd / cp / mv / rm / mkdir / chmod
   - 文件查看:cat / less / head / tail / grep / find
   - 系统信息:top / htop / df / free / uname
   - 网络工具:curl / wget / netstat / ss / ping

Credential Access

High
Category
Privilege Escalation
Confidence
60% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · tracks/deployment.md (reported line 68)May include surrounding context.

md
3. **docker-compose 多服务编排**
   - 编写 docker-compose.yml——"把多个容器的配置写在一个文件里,一键启动整个应用栈"
   - 常见编排:app + 数据库 + Redis
   - 环境变量管理——".env 文件存放配置,不要把密钥写死在 compose 文件里"
   - 常用命令:up / down / logs / exec / ps

⏸ **暂停点**:确认用户能用 Docker 成功构建镜像、启动容器后再继续。让用户实际操作一遍。

The database migration step lacks a warning about changes to existing database schemas and data.

Medium
Category
Output Handling
Confidence
92% confidence
Finding

The workflow instructs the agent to configure a database connection and run a migration without limiting the target to a disposable development database or warning about effects on existing data. The subsequent checkpoint confirms completion but does not disclose migration risks before the operation.

Content

No source excerpt is available for this finding.

Sudo/Root Execution

Medium
Category
Privilege Escalation
Confidence
70% confidence
Finding

Commands invoke sudo or root privileges. Verify this elevated access is necessary and justified.

Content

Scanner excerpt · tracks/deployment.md (reported line 40)May include surrounding context.

md
- 权限模型:rwx(读/写/执行)——"就像文件夹的访问权限:能看、能改、能运行"
   - chmod / chown 命令
   - 用户和用户组
   - sudo 的原理——"临时获得管理员权限,不是真的变成 root"

3. **SSH 连接与安全配置**
   - SSH 密钥登录配置——"密钥登录比密码安全得多,就像门禁卡比钥匙安全"

The SSH hardening workflow disables password login without warning about possible server lockout.

Medium
Category
Output Handling
Confidence
92% confidence
Finding

The workflow directs learners to disable SSH password login without first requiring a successful, independent key-authenticated login test or a recovery path. Although key setup precedes this step, an incorrect configuration can lock the learner out of the server; the later connectivity checkpoint comes too late to prevent that availability failure.

Content

No source excerpt is available for this finding.