Trust Assessment
core-vitals-fixer received a trust score of 65/100, placing it in the Caution category. This skill has some security considerations that users should review before deployment.
SkillShield's automated analysis identified 2 findings: 1 critical, 0 high, 1 medium, and 0 low severity. Key findings include Unpinned npm dependency version, Second-order Prompt Injection via User-Provided Code.
The analysis covered 4 layers: Manifest Analysis, Static Code Analysis, Dependency Graph, LLM Behavioral Safety. All layers scored 70 or above, reflecting consistent security practices.
Last analyzed on February 12, 2026 (commit 13146e6a). SkillShield performs automated 4-layer security analysis on AI skills and MCP servers.
Layer Breakdown
Behavioral Risk Signals
Security Findings2
| Severity | Finding | Layer | Location | |
|---|---|---|---|---|
| CRITICAL | Second-order Prompt Injection via User-Provided Code The skill reads arbitrary code files from a user-specified directory and directly concatenates their content into the 'user' message sent to the OpenAI API. This allows an attacker to embed malicious instructions within their source code (e.g., in comments or string literals) which could manipulate the LLM's behavior, override its system prompt, or attempt to exfiltrate information (e.g., the system prompt itself, or other contextual data the LLM might have access to). The LLM is instructed to analyze the code, but if the code contains instructions like 'ignore previous instructions', it could lead to unintended outputs. Implement robust sanitization or a more secure prompting strategy. Instead of directly injecting raw user code into the 'user' message, consider: 1) Using a dedicated tool call or function calling mechanism to pass code snippets as structured arguments. 2) Employing a 'sandwich' prompt structure where the user's code is encapsulated between clear delimiters and explicit instructions to treat it as data, not instructions. 3) Filtering or escaping potentially malicious keywords or patterns within the code before sending it to the LLM. 4) Instructing the LLM to strictly adhere to its system prompt and ignore any conflicting instructions found in the user content. | LLM | src/index.ts:25 | |
| MEDIUM | Unpinned npm dependency version Dependency 'commander' is not pinned to an exact version ('^12.1.0'). Pin dependencies to exact versions to reduce drift and supply-chain risk. | Dependencies | skills/lxgicstudios/core-vitals-fix/package.json |
Scan History
Embed Code
[](https://skillshield.io/report/0d7196637d064b59)
Powered by SkillShield