autosave

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 calls the autosave backend to fetch and parse user-provided public share content (e.g., pan.quark.cn links) — see docs/actions-tasks.md and docs/actions-detail.md and the implementation in scripts/api.cjs (task_suggestions, get_share_detail, fetchDirectoryTree) — so untrusted, user-generated share metadata and file lists are ingested and used to validate, sort, autofill tasks, and drive subsequent actions.