mcp-builder
Pass
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill describes a process that involves fetching documentation from external sources and exploring target server content, creating a vulnerability surface for indirect prompt injection. \n
- Ingestion points: SKILL.md (fetching documentation from modelcontextprotocol.io and github.com), reference/evaluation.md (inspecting content via read-only tool calls). \n
- Boundary markers: No specific boundary markers or 'ignore' instructions are provided for handling fetched documentation. \n
- Capability inventory: File system access (writing project files), shell command execution (building and testing servers), and network access (WebFetch). \n
- Sanitization: No explicit sanitization of external content is recommended. This is considered low risk given the developer-centric nature of the skill.
- [SAFE]: The skill adopts best practices for credential management by instructing the use of environment variables and providing safe placeholders (e.g., 'your_api_key_here', 'ghp_xxx'). References to external documentation are confined to well-known, official organizations and services, which are documented neutrally and do not represent a security risk.
Audit Metadata