ClawHub
Back to skill

Security audit

volcengine-troubleshooting

Security checks across malware telemetry and agentic risk

Overview

This troubleshooting skill is mostly transparent and read-oriented, but some reference tables can steer an agent toward retrieving secrets or running disruptive diagnostic actions.

Review this skill before installing. If used, run it only with least-privilege VolcEngine credentials, avoid raw output sharing, and do not allow GetSecretValue, BatchGetSecretValue, ListKubeconfigs, DescribeSecretKey, Redis failover/shutdown tests, KMS decrypt/sign/data-key operations, or other sensitive actions unless you explicitly requested them and have a redaction/no-log plan.

SkillSpector

By NVIDIA
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 (13)

Description-Behavior Mismatch

Medium
Confidence
97% confidence
Finding
The matrix explicitly includes `BatchGetSecretValue` and `GetSecretValue` as acceptable troubleshooting evidence for a troubleshooting skill. That normalizes retrieval of plaintext secret material during diagnosis, which exceeds least-privilege troubleshooting and creates a direct path to sensitive data exposure if an agent or operator follows the guidance.

Description-Behavior Mismatch

Medium
Confidence
98% confidence
Finding
In the KMS section, `BatchGetSecretValue` and `GetSecretValue` are classified as routine evidence-gathering actions alongside descriptive APIs. In a security/KMS troubleshooting context this is more dangerous, not less, because the referenced APIs can disclose the very secrets the system is meant to protect.

Intent-Code Divergence

Medium
Confidence
96% confidence
Finding
This matrix explicitly classifies Redis `Test*` actions as read-only/diagnostic candidates, but failure-injection or shutdown/failover test APIs can change runtime state, availability, or topology. In a troubleshooting skill, that misclassification is especially dangerous because an agent may auto-select these actions during diagnosis and trigger service disruption without explicit user approval.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The documented troubleshooting flow explicitly includes commands to retrieve instance user data and console output, which commonly contain bootstrap scripts, embedded secrets, tokens, credentials, or internal configuration. Presenting these commands without any warning or redaction guidance increases the chance that operators will expose sensitive data in logs, tickets, or shared troubleshooting channels.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The cloud assistant history and invocation result commands can reveal previously executed commands, their parameters, output, execution users, and possibly secrets printed by scripts or operational tooling. Because the skill is specifically for troubleshooting production-like compute environments, omission of a warning makes accidental disclosure during support workflows more likely and increases the sensitivity of the recommended data collection.

Missing User Warnings

Medium
Confidence
90% confidence
Finding
The matrix explicitly recommends `DescribeUserData` as a troubleshooting read action, but user-data commonly contains bootstrap secrets, API tokens, SSH material, or internal endpoints. Although the document says not to write credentials to logs, it does not warn that this specific API can return highly sensitive plaintext, making accidental disclosure during troubleshooting likely.

Missing User Warnings

Medium
Confidence
94% confidence
Finding
These entries include secret- and key-oriented read operations such as `ListE2BAPIKeys` and adjacent credential-related actions, but the matrix presents them as normal evidence-gathering steps without emphasizing that returned values may expose live authentication material. In a troubleshooting skill, that omission raises the risk that an agent will retrieve and surface credentials unnecessarily.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
`DescribeSecretKey` is a direct secret-retrieval action, yet the matrix lacks a prominent warning that its output may contain credential material. Including it as standard troubleshooting evidence can normalize unsafe collection and disclosure of secrets during routine support flows.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The document lists secret-retrieval actions without any nearby warning that these APIs may return sensitive plaintext and must not be echoed into chats, logs, or reports. In an agent skill, omission of that warning materially increases the chance of accidental disclosure during otherwise routine troubleshooting.

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The KMS coverage section presents secret-value retrieval as standard troubleshooting evidence but does not warn about plaintext disclosure risk. Because this is a KMS-focused skill, readers may over-trust the listed actions and inadvertently expose secrets while diagnosing unrelated access or configuration problems.

Credential Access

High
Category
Privilege Escalation
Content
| `ListAddons` | `vke` | 资源存在性、状态、配置或诊断证据 |
| `ListClusters` | `vke` | 资源状态、实例/节点/集群运行态证据 |
| `ListInstanceTypeLabels` | `vke` | 资源状态、实例/节点/集群运行态证据 |
| `ListKubeconfigs` | `vke` | 资源存在性、状态、配置或诊断证据 |
| `ListNodePools` | `vke` | 资源状态、实例/节点/集群运行态证据 |
| `ListNodes` | `vke` | 资源状态、实例/节点/集群运行态证据 |
| `ListPermissions` | `vke` | 身份、权限、密钥或授权证据 |
Confidence
96% confidence
Finding
Kubeconfig

Credential Access

High
Category
Privilege Escalation
Content
| `BatchGetSecretValue` | `kms` | 密钥、安全策略、风险或告警证据 |
| `DescribeCustomKeyStores` | `kms` | 密钥、安全策略、风险或告警证据 |
| `DescribeKey` | `kms` | 密钥、安全策略、风险或告警证据 |
| `DescribeKeyrings` | `kms` | 密钥、安全策略、风险或告警证据 |
| `DescribeKeys` | `kms` | 密钥、安全策略、风险或告警证据 |
| `DescribeRegions` | `kms` | 资源存在性、状态、配置或诊断证据 |
| `DescribeSecret` | `kms` | 密钥、安全策略、风险或告警证据 |
Confidence
92% confidence
Finding
Keyring

Credential Access

High
Category
Privilege Escalation
Content
| `GetPublicKey` | `kms` | 密钥、安全策略、风险或告警证据 |
| `GetSecretValue` | `kms` | 密钥、安全策略、风险或告警证据 |
| `ListTagsForResources` | `kms` | 资源存在性、状态、配置或诊断证据 |
| `QueryKeyring` | `kms` | 密钥、安全策略、风险或告警证据 |
| `VerifyMac` | `kms` | 资源存在性、状态、配置或诊断证据 |

### 攻击面管理
Confidence
99% confidence
Finding
Keyring

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.