Back to skill

Security audit

IPaaS VitePress Content Automation

Security checks for vulnerabilities and agentic risk

Overview

The skill is mostly aligned with generating and publishing VitePress documents, but it needs Review because its SSH deployment path is under-scoped and can create persistent remote access.

Install only if you are comfortable with a skill that can modify local site files and publish them over SSH. Use a dedicated low-privilege deployment account, verify SSH host-key fingerprints out of band, restrict the authorized key and remote write directory, validate SERVER_PORT and REMOTE_DIR yourself, and require an explicit manual confirmation before running deployment.

Vulnerability Patterns
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
  • System PersistenceInstalls backdoors, hooks, services, or scheduled tasks that survive the run
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
Findings (5)

T06 · System Persistence

Error
Location
README.md:4
Finding

Persistent Remote Access Through SSH Authorized-Key Installation

Content
View full analysis
> ~/.ssh/known_hosts` 2. **免密配置**:必须完成 `ssh-copy-id` 授权。 ``` ### Technical Analysis The documented setup requires the user to run `ssh-copy-id`. This command normally adds a public key to the remote account's `~/.ssh/authorized_keys` file. The resulting authorization persists after the deployment process terminates and continues to permit authentication in future sessions. Although public-key authentication is appropriate for automated deployment, installing a persistent authorized key exceeds a single deployment run unless the key is separately managed, tightly restricted, and subject to a documented revocation process. Possession or later compromise of the corresponding private key would provide durable access to the remote deployment account. ### Attack Path 1. The user follows the mandatory setup instructions and runs `ssh-copy-id`. 2. The command adds the user's public key to the remote deployment account's `authorized_keys`. 3. The authorization remains installed after the Skill finishes. 4. An attacker later obtains the corresponding private key or access to the system holding it. 5. The attacker authenticates to the remote account without repeating the setup process. 6. The attacker exercises all permissions available to that deployment account. ### Impact Assessment A successful attacker obtains persistent access to the remote deployment account. The accessible scope includes files, commands, and services available to that account. The impact becomes substantially greater if the account is privileged, has broad write access, or can escalate privileges on the remote server. ]]>
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Error
Location
script/deploy.sh:13
Finding

SSH Option Injection Through Unvalidated SERVER_PORT

Content
View full analysis
Remediation
View remediation
65535 )); then echo "Error: SERVER_PORT must be an integer from 1 to 65535." >&2 exit 1 fi ``` - Reject all whitespace, option prefixes, control characters, and nonnumeric input. - Avoid constructing commands by concatenating untrusted values into command strings. - Validate `SERVER_IP` and `REMOTE_USER` against strict allowlists appropriate to their expected formats. - Run deployment under a dedicated, minimally privileged local identity. - Add tests covering malicious values containing spaces, quotes, newlines, and SSH options. ]]>

T05 · Unauthorized Access and Privilege Escalation

Error
Location
script/deploy.sh:23
Finding

Remote Path Blacklist Can Be Bypassed

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
README.md:4
Finding

Unauthenticated SSH Host-Key Enrollment Guidance

Content
View full analysis
> ~/.ssh/known_hosts` ``` ### Technical Analysis `ssh-keyscan` retrieves the host key presented over the network but does not authenticate that the returned key belongs to the intended server. Blindly appending its output to `~/.ssh/known_hosts` converts an unauthenticated network response into a persistent trust decision. If an attacker can intercept or redirect the initial scan, the attacker's host key may be enrolled. Subsequent SSH and rsync connections can then trust the impersonating host because its key matches the locally stored entry. ### Attack Path 1. The user follows the README before the first deployment. 2. A network attacker intercepts or redirects the `ssh-keyscan` request. 3. The attacker returns a host key under their control. 4. The user appends that key to `~/.ssh/known_hosts` without independent fingerprint verification. 5. A later deployment connects to infrastructure controlled or intercepted by the attacker. 6. SSH accepts the enrolled key, allowing server impersonation. ### Impact Assessment An attacker may impersonate the deployment server, receive the generated site content, observe deployment metadata, and misdirect deployment operations. SSH public-key authentication does not normally disclose the private key, but successful host impersonation still breaks destination authenticity and can expose transferred content. ]]>
Remediation
View remediation

T08 · Insecure Dependencies

Warning
Location
script/deploy.sh:29
Finding

Uncontrolled Dependency Installation During Deployment

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (6)

Credential Access

High
Category
Privilege Escalation
Confidence
90% confidence
Finding

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

Content

Scanner excerpt · README.md (reported line 6)May include surrounding context.

md
## 核心安全要求
为通过安装校验,本工具包遵循高安全标准:
1. **主机密钥校验**:脚本不再自动接受未知主机。请在首次运行前执行:
   `ssh-keyscan -p [端口] [服务器IP] >> ~/.ssh/known_hosts`
2. **免密配置**:必须完成 `ssh-copy-id` 授权。
3. **路径限制**:禁止向系统根目录同步。

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The trigger condition is overly broad: any request to write or publish an integration solution could activate a workflow that reads templates, writes documentation files, modifies site configuration, and initiates deployment. In a user-invocable skill with shell access, broad activation increases the chance of unintended execution of state-changing actions without sufficiently explicit user intent.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The workflow performs impactful actions—writing content, modifying VitePress configuration, and executing a deployment shell script—but the skill does not clearly warn that it will change local files and publish to a remote server. Because it also uses shell, rsync, and ssh-based deployment, a user may unknowingly trigger irreversible or externally visible changes.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

This code file contains natural-language instructions and comments exclusively in Chinese, including operational guidance about path detection and environment variable requirements. The policy only permits a fixed language when the skill offers user choice or clearly justifies the locale constraint, which is not present here.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The file states that product, inventory, shipment, and return data are synchronized automatically with no manual intervention. It does not warn users that the integration can modify live storefront inventory and create ERP sales/return documents, which could affect user data and system integrity if misconfigured.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
93% confidence
Finding

The skill documentation forces a specific language/locale for user-facing instructions, which can violate language-choice policy when no opt-in or alternative is provided. The file contains operational and safety guidance entirely in Chinese, with no indication that this is a region-specific or language-specific skill.

Content

No source excerpt is available for this finding.