atlassian-mcp
Pass
Audited by Gen Agent Trust Hub on Jun 14, 2026
Risk Level: SAFE
Full Analysis
- [EXTERNAL_DOWNLOADS]: The documentation references official and community-supported MCP servers available on standard registries like npm and PyPI. This includes the trusted '@anthropic/mcp-atlassian' package and common community alternatives such as 'mcp-atlassian'.
- [COMMAND_EXECUTION]: Configuration examples include the use of 'npx' and 'uvx' to run MCP server executables, which is standard practice for local tool integration.
- [PROMPT_INJECTION]: The skill provides guidance on mitigating indirect prompt injection by recommending sanitization functions and the use of 'maxResults=1' probes to validate queries before full execution. It also explicitly forbids hardcoding API tokens or secrets.
Audit Metadata