Back to skill

Security audit

测试左移

Security checks for vulnerabilities and agentic risk

Overview

The skill is a coherent QA testing guide, but it repeatedly recommends an unpinned remote installation that would add a larger external skill set outside the reviewed artifact.

Install this skill only if you are comfortable with Chinese-language QA workflow guidance and review any project files before allowing Bash-based actions. Do not run the recommended `npx skills add Kokxi/qa-test-skills` command unless you separately trust and verify that external collection, ideally using a pinned commit or signed release.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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
Findings (1)

T08 · Insecure Dependencies

Warning
Location
SKILL.md:8
Finding
Unpinned Third-Party Skill Installation Creates a Supply-Chain Risk## Vulnerability Details **File Location**: `SKILL.md`, line 8; repeated at line 56 **Vulnerability Type**: Unpinned and unverified third-party dependency installation **Risk Level**: Medium **Vulnerable code at line 8:** ```yaml description: >- 当项目还在需求阶段或者开发正在写代码时使用此技能——这时候介入能花最小的成本避免最多的缺陷。从需求可测试性评审(需求模糊/矛盾/不可测)、开发阶段测试设计(单元测试/接口契约/测试桩)和技术方案评审(影响面分析/风险识别)三个维度提前发现缺陷。越早发现 Bug 修复成本越低——需求阶段的 Bug 修复成本是线上阶段的 1/100。输出左移检查清单和阶段性介入记录。 本技能属于 QA Test Skills 技能集(49 个技能之一),完整工作流体验需安装全套:npx skills add Kokxi/qa-test-skills ``` **Repeated installation instruction at line 56:** ```markdown > ⚠️ 本技能单独使用效果有限,建议配合完整技能集(12 步工作流)使用。安装:npx skills add Kokxi/qa-test-skills ``` ### Technical Analysis The Skill recommends running `npx skills add Kokxi/qa-test-skills`, but it does not pin the referenced repository to an immutable commit, audited release, or verified integrity digest. Consequently, the content installed by this command may differ from the content that existed when this Skill was audited. The command crosses two mutable trust boundaries: `npx` may dynamically resolve and execute package tooling, and that tooling then retrieves a third-party skill collection identified only by its repository name. If either dependency source is compromised or later changed maliciously, the installation can introduce unaudited code or agent instructions. The command is documentation rather than an automatically executed payload, and the reviewed project does not establish that the current upstream dependency is malicious. The vulnerability is therefore a conditional supply-chain exposure requiring a user or agent to execute the recommended command. ### Attack Path 1. An attacker compromises the referenced repository, a maintainer account, or a package involved in resolving the `npx skills` command. 2. The attacker adds malicious scripts, unsafe lifecycle behavior, or hostile agent instructions to the mutable upstream content. 3. A user or a ...[truncated 1277 chars]
Remediation
## Remediation Suggestions 1. Remove the third-party installation recommendation if the external collection is not necessary for the Skill's core operation. 2. If installation is retained, pin the source to an immutable, reviewed commit hash or cryptographically signed release rather than a mutable repository reference. 3. Verify downloaded content using a trusted checksum or signature before installation. 4. Pin the `npx` package and version used to perform installation, and use a trusted registry with lockfile and integrity verification where supported. 5. Require explicit user confirmation before retrieving or installing external content; the Skill must not run the command automatically. 6. Document all files, scripts, lifecycle hooks, permissions, and network operations introduced by the dependency. 7. Review the pinned upstream content independently and repeat the security audit before changing the pinned revision. 8. Perform installation in a restricted environment using least privilege, limited filesystem access, no unnecessary secrets, and constrained network access.
Vulnerability Patterns
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • 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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (5)

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding
MCP server references in the skill manifest without version pinning are a rug-pull risk.

Natural-Language Policy Violations

Medium
Confidence
88% confidence
Finding
The display name and all user-facing instructional content are written in Chinese, but the file does not state that the skill is region-specific or that users may choose another language. Under the policy, a fixed language presentation without opt-in or justification is a natural-language policy concern.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding
The skill instructs users to run `npx skills add Kokxi/qa-test-skills` without pinning an exact package/version. This allows whatever package version is current at execution time to be fetched and executed, creating a supply-chain risk if the package is compromised, replaced, or updated with malicious code.

Vague Triggers

Medium
Confidence
95% confidence
Finding
The invocation description includes generic phrases like “左移”, “提前测试”, and especially the condition “项目早期需要介入测试时”, which overlap with common discussion about software process rather than a clearly delimited skill trigger. The file does not provide exclusion conditions or negative examples to clarify when the skill should not activate.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding
This line again recommends executing `npx skills add Kokxi/qa-test-skills` without a pinned version. Repeated unpinned remote execution increases the chance that users will run mutable third-party code, which is a classic supply-chain exposure.

Static analysis

No suspicious patterns detected.