review-gitlab

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

• Third-party content exposure detected (high risk: 0.95). The skill’s required runtime workflow fetches MR diff content and existing discussion notes via the bundled CLI (get_mr_diff → /merge_requests/:iid/diffs and get_mr_comments → /merge_requests/:iid/discussions), which are outsider-authored free text (other contributors’ code/comments) that is then ingested into the agent’s LLM context for review.

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

• Potentially malicious external URL detected (high risk: 0.80). The bundled CLI makes runtime requests to the GitLab API (e.g. https://gitlab.com/api/v4 via MR_MCP_GITLAB_HOST) to fetch merge request diffs and comments which are required by the skill and directly drive the agent's review prompts and posted outputs, so this is a runtime external dependency that can control agent behavior.

MEDIUM W021: Hidden or invisible Unicode characters detected (potential obfuscation or prompt injection).

• Hidden Unicode characters detected (1 type(s) found)