Commit Graph
7 Commits
Author SHA1 Message Date
benjamin b92bbf0299 fix: 过滤入口拒绝日志并强化鉴权边界 2026-07-18 00:11:18 +08:00
mt21625457 ac685ccaf5 feat(security-audit): persist full prompts on audit events and polish event review UI
- Add prompt_audit_events.full_prompt (migration 182) so admins can review
  the exact unredacted prompt that triggered a finding; blocking mode writes
  it from the snapshot, async mode reconstructs it from the Redis scan
  payload so jobs rows stay redaction-only
- Event detail API returns full_prompt (list endpoint stays lean); text is
  NUL-stripped and capped at 65536 runes
- Detail dialog shows the full prompt in a scrollable pane with fallback to
  the legacy redacted preview; page copy updated to match the new behavior
- Rework filter deletion into a dedicated dialog with time-range presets and
  criteria-change preview invalidation; localize decision/risk/category
  labels across the events workspace
- Fix pre-existing i18n message-compile spec by declaring the
  @intlify/message-compiler dev dependency
2026-07-17 14:38:10 +08:00
mt21625457 7ed4e7e5e3 fix(security-audit): isolate latest user prompt chunk 2026-07-17 12:12:41 +08:00
mt21625457andCursor 18e698bed6 feat(security-audit): allow admin-managed audit node targets and polish pool UI
Let admins configure private/intranet Guard endpoints without destination-class blocking, and fix prompt-audit switch layout so thumbs and labels no longer overlap.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-07-17 11:45:14 +08:00
mt21625457andCursor df9d9e2e40 fix(security-audit): harden role scan, startup, probe, and localhost dial
Scan client-injected assistant/tool/model turns, fail closed when config cannot
be trusted after startup or stale invalidation, reuse probe tokens only for the
same base URL, and restrict localhost dials to loopback addresses.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-07-17 09:00:14 +08:00
mt21625457andCursor 0f7f8a317e fix(security-audit): close prompt-audit bypass and privacy gaps
Stop WebSocket follow-up turns from reusing a request-wide audit cache, scan
client-controlled instruction fields, fail closed on stale weaker configs, and
tighten preview/SSRF controls including persisted request stage.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-07-17 08:46:57 +08:00
mt21625457 d11bdb13f5 feat(security-audit): add OpenAI-compatible prompt auditing 2026-07-17 00:39:39 +08:00