tool-list
Fail
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill documentation suggests installing the
unity-mcp-clipackage globally vianpm install -g unity-mcp-cli. This downloads code from a public registry that is not associated with a verified trusted organization or well-known service provided in the analysis guidelines. - [REMOTE_CODE_EXECUTION]: The skill suggests using
npx unity-mcp-cli, which automatically downloads and executes the package's code in the current environment. This represents a potential execution of unverified remote code. - [INDIRECT_PROMPT_INJECTION]: The skill accepts user-provided input through the
regexSearchparameter, which is passed as a string to theunity-mcp-clitool. This creates a surface where malicious data could influence tool behavior. - Ingestion points: Input parameters defined in the
SKILL.mdfrontmatter and JSON schema. - Boundary markers: No delimiters or instructions are used to separate the regex input from the command context.
- Capability inventory: The skill executes shell commands (
unity-mcp-cli) using the provided inputs. - Sanitization: There is no evidence of input validation or regex sanitization to prevent potential command injection or resource exhaustion (ReDoS).
Recommendations
- AI detected serious security threats
Audit Metadata