external-gitcode-ascend-msverl-daily-regression-triage

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

• Third-party content exposure detected (high risk: 0.80). The skill's required workflow (SKILL.md) and scripts (sync_repos.py, list_recent_commits.py, rank_candidate_commits.py) explicitly clone and read public repositories (https://github.com/verl-project/verl.git and https://gitcode.com/Ascend/MindSpeed.git) and then parse commit messages/files to rank suspects, so untrusted, user-generated repository content is fetched and directly influences triage decisions.