Back to skill

Security audit

测试领导力

Security checks for vulnerabilities and agentic risk

Overview

The skill itself is a coherent QA leadership guide, but it recommends an unpinned command that could install a broad third-party skill collection and change the user's agent environment.

Install this only if you want a Chinese-language QA leadership skill and are comfortable reviewing it manually. Do not run the suggested `npx skills add Kokxi/qa-test-skills` command unless you first verify the exact package, version or commit, source, and contents of the broader 49-skill collection, because that command can change what skills your agent uses in future sessions.

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 Supply-Chain Risk## Vulnerability Details **File Location**: `SKILL.md`, lines 8 and 49 **Vulnerability Type**: Unpinned third-party dependency installation **Risk Level**: Medium ### Vulnerable Code Line 8: ```shell npx skills add Kokxi/qa-test-skills ``` Line 49: ```shell npx skills add Kokxi/qa-test-skills ``` ### Technical Analysis The skill twice recommends using `npx` to install a third-party skill collection without specifying an immutable installer version, repository commit, checksum, or signature. Both the package resolved by `npx` and the referenced skill collection may therefore change after this skill has been reviewed. If either upstream distribution channel is compromised or publishes a malicious update, users following the documented command could retrieve and process content different from the version originally assessed. The project provides no integrity-verification procedure or evidence that the retrieved package and skill files have been independently reviewed. The command is presented as a recommendation rather than being automatically executed by the audited skill. Exploitation therefore requires a user or agent to follow the installation instruction. ### Attack Path 1. An attacker compromises the package resolved by `npx`, the referenced repository, or an associated publishing account. 2. The attacker publishes a modified installer or skill collection under the same mutable package or repository identifier. 3. A user or agent follows the command documented in `SKILL.md`. 4. `npx` resolves the currently available installer, which may differ from the reviewed version. 5. The installer retrieves or installs the attacker-controlled skill content. 6. Malicious instructions, configuration, or executable lifecycle behavior may then run with the privileges of the invoking user or influence later agent sessions. ### Impact Assessment Successful exploitation could modify the user's installed Agent ski ...[truncated 482 chars]
Remediation
## Remediation Suggestions 1. Pin the `npx` installer to a specific reviewed version rather than relying on the latest registry resolution. 2. Pin the external skill collection to an immutable release or full commit hash. 3. Publish expected cryptographic checksums or signatures and require verification before installation. 4. Use a trusted package registry and enable provenance or signature verification where supported. 5. Review the downloaded package, lifecycle scripts, and skill files before allowing installation. 6. Run installation in a sandbox or least-privileged environment without access to credentials or sensitive project files. 7. Document the exact versions that were security-reviewed and establish a controlled process for reviewing upgrades.
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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
74% confidence
Finding
The manifest-level finding indicates the skill package references an MCP-related installer command without version pinning. Even if the actual executable string appears in later fields, the practical issue remains a supply-chain exposure: consumers of the manifest may trust and automate installation based on metadata that does not constrain the version. That can lead to execution of altered code outside the originally reviewed skill behavior.

Natural-Language Policy Violations

Medium
Confidence
89% confidence
Finding
The skill is presented entirely in Chinese, including the display name and usage instructions, with no indication that users may request another language. This can violate language/locale policy when the skill effectively assumes a fixed output language without opt-in or documented regional justification.

Rp1

Medium
Category
MCP Rug Pull
Confidence
92% confidence
Finding
The skill metadata instructs users to install a package via 'npx skills add Kokxi/qa-test-skills' without pinning an exact version. This makes executions non-reproducible and exposes users to supply-chain risk if a newer compromised package or dependency is published under the same name. In skill content, install commands are more dangerous because they may be copied and run directly by operators.

Vague Triggers

Medium
Confidence
94% confidence
Finding
The `when_to_use` field includes very common terms such as "目标", "招聘", and "面试", plus broad conditions like "需要管理测试团队", without clear scope boundaries or exclusion examples. These phrases can overlap with ordinary conversations and may cause unintended invocation outside the intended QA test leadership context.

Rp1

Medium
Category
MCP Rug Pull
Confidence
92% confidence
Finding
This line repeats an unpinned 'npx skills add Kokxi/qa-test-skills' installation command in the body text. Recommending unpinned package execution allows the fetched code to change over time, increasing the chance of accidental execution of malicious or unreviewed updates. Because it is presented as a setup step, users may run it without additional scrutiny.

Static analysis

No suspicious patterns detected.