Back to skill

Security audit

测试批判性思维

Security checks for vulnerabilities and agentic risk

Overview

The skill itself is a QA thinking guide, but it tells users to run an unpinned command that can install a larger external skill set.

Review the skill before installing the broader QA Test Skills set. The inspected skill is read-only QA guidance, but avoid running the recommended unpinned `npx` install command unless you trust the source and are comfortable installing a larger mutable skill collection; prefer a pinned commit or verified 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`, lines 8 and 50 **Vulnerability Type**: Unpinned third-party dependency and mutable repository reference **Risk Level**: Medium ### Vulnerable Code The installation command appears twice: ```shell npx skills add Kokxi/qa-test-skills ``` ### Technical Analysis The documented command invokes the `skills` package through `npx` without specifying an exact, reviewed package version. It also identifies the external skill collection using the mutable repository reference `Kokxi/qa-test-skills`, rather than an immutable commit identifier with integrity verification. Consequently, the software retrieved when a user follows this recommendation may differ from the content that was available when this project was audited. Compromise of the package registry account, the referenced repository, its maintainers, or the dependency-resolution path could cause modified content to be downloaded. The command is documentation and is not executed automatically by the audited skill. Exploitation therefore requires a user or automation process to follow the installation recommendation. ### Attack Path 1. An attacker compromises or otherwise gains control over the unpinned `skills` package, the referenced repository, or an associated distribution account. 2. The attacker publishes a modified package or changes content resolved through the mutable repository reference. 3. A user follows the installation guidance and runs `npx skills add Kokxi/qa-test-skills`. 4. `npx` resolves and runs the available CLI package, which retrieves or installs the current external skill content. 5. Malicious package behavior, installation hooks, scripts, or skill instructions execute or become available in the user's environment. ### Impact Assessment If the resolved CLI package executes malicious code, it could operate with the privileges of the user running `npx`. Depending on the user's permissions and ...[truncated 528 chars]
Remediation
## Remediation Suggestions 1. Pin the `skills` CLI to an exact reviewed version, for example by using an explicit package version rather than the latest registry resolution. 2. Pin the external skill collection to a verified immutable commit or signed release rather than a mutable repository name or branch. 3. Publish and verify cryptographic hashes or signatures for all installed files. 4. Review downloaded skill files and scripts before loading or executing them. 5. Disable lifecycle scripts during package retrieval where supported, unless those scripts have been explicitly reviewed. 6. Run installation in a sandbox or least-privileged environment without production credentials or access to sensitive files. 7. Use an approved internal registry or vendored, reviewed copy of the dependency where possible. 8. Update both occurrences of the installation guidance so users are not directed to the unpinned command.
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 skill metadata and instructional content are entirely in Chinese, and there is no statement that the user may choose another language or locale. This can violate language/locale policy when a skill implicitly forces a specific language without opt-in or documented regional justification.

Rp1

Medium
Category
MCP Rug Pull
Confidence
91% confidence
Finding
The skill instructs users to run `npx skills add Kokxi/qa-test-skills` without pinning a specific version. Because `npx` resolves and executes the latest matching package/tooling at runtime, a compromised upstream release or typosquatted dependency could cause unreviewed code to run on the user's machine. In this skill context, the risk is real because the command is presented as installation guidance, making execution likely.

Vague Triggers

Medium
Confidence
96% confidence
Finding
The `when_to_use` field includes broad phrases like '你怎么看', '还有什么', and '这样够吗', which can match ordinary conversation rather than an explicit request for this skill. In agentic systems, overly broad triggers can cause unintended invocation, injecting this skill's instructions and shaping the assistant's behavior in contexts where the user did not ask for it. Because this is a testing skill that reframes analysis through skeptical prompts, the impact is workflow confusion and possible prompt-scope interference rather than direct code execution.

Rp1

Medium
Category
MCP Rug Pull
Confidence
91% confidence
Finding
This line repeats an unpinned `npx skills add Kokxi/qa-test-skills` installation command. Repetition increases the chance a user will execute it, and the lack of version pinning preserves supply-chain risk from future malicious or breaking upstream changes. The surrounding text does not mitigate this risk.

Static analysis

No suspicious patterns detected.