Back to skill

Security audit

vmware-aiops

Security checks for vulnerabilities and agentic risk

Overview

ClawScan rates the VMware management skill suspicious with medium confidence, citing concerns about its documented capabilities and setup, while noting that some referenced material was only partially inspected.

Treat this as a review warning rather than a finding of malicious intent. Before installing, consider whether the documented assistant configuration changes, broad VMware write operations, external model examples, and support-sharing guidance fit your needs. Coverage is incomplete: 8 of 11 components were scanned, and some references had obfuscated text or reached analyzer output limits.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (128)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The supplied commands modify a local assistant's project-rule configuration rather than perform VMware operations. Installing the skill document as project rules is not disclosed in the declared purpose and is not merely an implementation detail of the described VMware capabilities. The copied document's contents are not supplied, so no further behavior is inferred.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The commands modify local project rule configuration rather than perform VMware operations. Installing skill content as project rules is an undeclared capability affecting assistant behavior, not a supporting implementation detail of the described VMware management functions.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The supplied shell commands set up and launch a general-purpose AI coding assistant. They do not perform VMware operations or demonstrate a supporting implementation step for those operations. Referencing the VMware skill file as Aider conventions does not make the commands VMware lifecycle or diagnostic functionality.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
96% confidence
Finding

The supplied commands set up and invoke a model-backed coding tool; they do not perform any declared VMware operation. Loading the skill document as coding conventions does not make this a VMware management action. Although this appears in a reference guide and may be development tooling, its actual behavior is materially different from the declared purpose. The snippet does not establish credential theft, successful external communication, or file modification.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
96% confidence
Finding

The supplied commands set up and launch a local AI coding environment rather than perform VMware administration. Referencing a VMware skill file as Aider conventions does not itself make these operations VMware lifecycle or investigation functionality. The declared description does not establish this coding environment as a supporting dependency.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
96% confidence
Finding

The supplied commands configure and start an AI coding tool; they do not implement any declared VMware operation. Referencing the VMware skill file as conventions does not establish VMware functionality. The chunk shows no credential theft or specific data exfiltration, but its demonstrated purpose materially differs from the declared operational purpose.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
94% confidence
Finding

The supplied commands initiate an AI-assisted coding workflow, not a VMware management operation. Although the VMware skill file is passed as conventions, the chunk contains no VMware API calls or administrative actions establishing this as a supporting implementation detail. The API key shown is a placeholder; the snippet does not establish credential theft or infrastructure-data exfiltration.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
94% confidence
Finding

The supplied commands implement an AI-assisted coding workflow, not the declared VMware administration capabilities. They introduce external model-provider integration that is not described in the declared purpose. The API key values are placeholders, so this snippet does not establish disclosure of real credentials. This assessment applies only to the supplied chunk, not the rest of the skill.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
94% confidence
Finding

The supplied shell commands set up and run an AI coding environment rather than perform VMware administration. Referencing the VMware skill document as Aider conventions does not make these commands VMware operations. Although the chunk is documentation, its executable behavior is not an evident supporting implementation detail of the declared infrastructure-management capabilities.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
94% confidence
Finding

The description does not disclose authentication administration or serving an externally accessible MCP endpoint. These commands expose administrative and network-service capabilities beyond the described VMware operations and scanning daemon. TLS configuration itself is a supporting detail, but token administration and public-facing service hosting are substantive undeclared capabilities. The snippet consists of documentation examples, so it does not establish that these commands execute automatically.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
92% confidence
Finding

The supplied commands implement an AI-assisted coding workflow, not any declared VMware management operation. Although they reference the VMware skill document, that document is supplied as Aider conventions; this does not make the commands perform VMware administration. The snippet does not establish that infrastructure data or real credentials are transmitted.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
90% confidence
Finding

The chunk describes access-control administration and remote service hosting rather than the declared VMware operations. These are operator-facing capabilities not disclosed by the description; the optional scanning/TTL daemon does not establish that an externally accessible MCP service or token administration is included. The chunk contains documented shell commands, so it demonstrates these capabilities without establishing that they execute automatically.

Content

No source excerpt is available for this finding.

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 · SKILL.md (reported line 22)May include surrounding context.

md
SSWORD","VMWARE_<TARGET>_USERNAME","SLACK_WEBHOOK_URL","DISCORD_WEBHOOK_URL","VMWARE_AUDIT_APPROVED_BY"],"bins":["vmware-policy"]},"homepage":"https://github.com/vmware-skills/VMware-AIops","emoji":"🖥️","os":["macos","linux"]}}
compatibility: >
  vmware-policy auto-installed as Python dependency (provides @vmware_tool decorator and audit logging). All write operations audited to ~/.vmware/audit.db.
  Credentials: Each vCenter/ESXi target requires a per-target password env var in ~/.vmware-aiops/.env following the pattern VMWARE_<TARGET_NAME_UPPER>_PASSWORD. Passwords are never logged or echoed.
  Destructive operations: All write tools require explicit parameters and pass through the @vmware_tool decorator (policy check + audit + sanitize). Every MCP write tool annotated destructive (23 of 43: power-off, delete, migrate, snapshot revert/delete, guest exec/upload/provision, cluster delete/remove-host, TTL, Clean Slate, plan apply/rollback, host-network, DRS and alarm reset) takes one co

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 60)May include surrounding context.

md
Read before connecting an agent. Per-tool inventory: `references/capabilities.md`.

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 67)May include surrounding context.

md
TPS host files additionally require `VMWARE_MCP_FILE_ROOT`. Upload boundaries: `references/capabilities.md`.

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 69)May include surrounding context.

md
Backend resize limits: see `references/capabilities.md`.

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 227)May include surrounding context.

md
re list (complete by construction). Rationale, `total` semantics, error shape: `references/capabilities.md`.

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 129)May include surrounding context.

md
X slow / failing / down" issue, follow [`references/investigation-protocol.md`](references/investigation-protocol.md). It enforces four root-cause completeness

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 · SKILL.md (reported line 316)May include surrounding context.

bash
uv tool install vmware-aiops==2.5.3
mkdir -p ~/.vmware-aiops
vmware-aiops init  # generates config.yaml and .env templates
chmod 600 ~/.vmware-aiops/.env

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 · references/detailed-guide-CN.md (reported line 94)May include surrounding context.

bash
# 使用模板创建 .env 文件
cp .env.example ~/.vmware-aiops/.env

# 编辑并填入真实密码
# 然后锁定文件权限(仅所有者可读写)

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 · references/setup-guide.md (reported line 47)May include surrounding context.

3. Configure

mkdir -p ~/.vmware-aiops vmware-aiops init # generates config.yaml and .env templates chmod 600 ~/.vmware-aiops/.env

Edit ~/.vmware-aiops/config.yaml and .env with your target details

text

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 · references/setup-guide.md (reported line 49)May include surrounding context.

mkdir -p ~/.vmware-aiops vmware-aiops init # generates config.yaml and .env templates chmod 600 ~/.vmware-aiops/.env

Edit ~/.vmware-aiops/config.yaml and .env with your target details

text

### Declare `environment:` on each target

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 · SKILL.md (reported line 317)May include surrounding context.

uv tool install vmware-aiops==2.5.3 mkdir -p ~/.vmware-aiops vmware-aiops init # generates config.yaml and .env templates chmod 600 ~/.vmware-aiops/.env

text

> Full setup guide, security details, and AI platform compatibility: see `references/setup-guide.md`

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 · references/detailed-guide-CN.md (reported line 93)May include surrounding context.

通过 .env 文件设置密码(推荐):

bash
# 使用模板创建 .env 文件
cp .env.example ~/.vmware-aiops/.env

# 编辑并填入真实密码

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 · references/detailed-guide-CN.md (reported line 98)May include surrounding context.

编辑并填入真实密码

然后锁定文件权限(仅所有者可读写)

chmod 600 ~/.vmware-aiops/.env

text

> **安全提示**:推荐使用 `.env` 文件而非命令行 `export`,避免密码出现在 shell 历史记录中。`.env` 文件必须设置 `chmod 600`(仅所有者可读写)。