gitlab-automation

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). In the GitLab “Manage Issues” required tool sequence, the workflow calls GITLAB_LIST_PROJECT_ISSUES with a free-text search parameter, which causes the LLM to ingest GitLab issue title/description text authored by GitLab users (outsiders) when listing/filtering issues.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill requires adding and calling the external MCP endpoint https://rube.app/mcp at runtime (via RUBE_SEARCH_TOOLS / RUBE_MANAGE_CONNECTIONS) which fetches tool schemas that can directly influence agent behavior and prompts, so this is a runtime dependency that controls the agent.