atlassian-rovo
Pass
Audited by Gen Agent Trust Hub on Mar 8, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: Fetches the mcp-remote proxy tool from the npm registry to establish a connection with Atlassian's official MCP endpoint at https://mcp.atlassian.com/v1/mcp.- [COMMAND_EXECUTION]: Instructs the user to run shell commands using npx to initialize the MCP server and complete interactive OAuth-based authentication for Jira and Confluence access.- [PROMPT_INJECTION]: The autonomous Resume protocol in Phase 3 creates a surface for indirect prompt injection by reading project status and worklogs from Jira issues and Confluence pages to determine subsequent agent team actions.
- Ingestion points: Uses atlassian:getJiraIssue, atlassian:getConfluencePage, and atlassian:searchJiraIssuesUsingJql to ingest data in phase-resume.md.
- Boundary markers: The workflow employs structured markdown templates for project plans and Epics, but lacks explicit delimiters or instructions to ignore embedded commands in the processed data.
- Capability inventory: The skill possesses extensive capabilities across all scripts including TeamCreate, TaskCreate, SendMessage, createJiraIssue, editJiraIssue, createConfluencePage, and updateConfluencePage.
- Sanitization: There is no evidence of sanitization or validation logic for the content retrieved from Atlassian platforms before it is interpolated into agent prompts.
Audit Metadata