confluence-expert
Pass
Audited by Gen Agent Trust Hub on Nov 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [SAFE]: The skill consists of markdown documentation and local Python scripts that perform static analysis of Confluence metadata. All scripts rely on standard library modules and do not require administrative privileges or network access.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from Confluence space exports. 1. Ingestion points: The three Python scripts in
scripts/read titles and labels from JSON files. 2. Boundary markers: Absent in the script-generated reports. 3. Capability inventory: The skill utilizes Atlassian Rovo MCP tools for content management and search. 4. Sanitization: JSON fields are parsed but not sanitized before report generation.\n- [EXTERNAL_DOWNLOADS]: The playbook references the official Atlassian Rovo MCP server (mcp.atlassian.com) for Cloud integration. As this targets a well-known service, the references are documented neutrally.
Audit Metadata