新增技能目录重新扫描接口与分页,关掉技能安全扫描的配置开关

- 后端新增 POST /api/skills/reconcile-custom:扫描 custom 目录,把磁盘存在但 DB 无记录的技能登记到调用者名下(published=false),附 4 个单元测试
- 前端技能列表加"重新扫描技能目录"按钮,列表分页每页 10 条
- SkillsConfig 新增 security_scan.enabled(默认 true)用于关闭 LLM 安全扫描;scan_skill_content 在开关关闭时直接放行
- 智能体卡片/对话页长名 truncate 显示 + loading 占位

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This commit is contained in:
吉鑫升andClaude Opus 4.7 committed 2026-05-22 16:22:42 +08:00
1 parent b6a2c9336e
commit 80fc2a9240
14 files changed
+366 -11

No files matched your search

@@ -39,6 +39,12 @@ def _extract_json_object(raw: str) -> dict | None:
async def scan_skill_content(content: str, *, executable: bool = False, location: str = SKILL_MD_FILE, app_config: AppConfig | None = None) -> ScanResult:
"""Screen skill content before it is written to disk."""
config = app_config or get_app_config()
# Operator-level kill switch: when ``skills.security_scan.enabled`` is false,
# bypass the LLM round-trip entirely and let the content through.
if not config.skills.security_scan.enabled:
return ScanResult("allow", "Security scan disabled by config.")
rubric = (
"You are a security reviewer for AI agent skills. "
"Classify the content as allow, warn, or block. "
@@ -49,7 +55,6 @@ async def scan_skill_content(content: str, *, executable: bool = False, location
prompt = f"Location: {location}\nExecutable: {str(executable).lower()}\n\nReview this content:\n-----\n{content}\n-----"
try:
config = app_config or get_app_config()
# 安全审核复用 curator 的评审模型配置(skill_evolution.curator.model_name):
# 二者用同一个模型。未配置时 model_name 为 None,create_chat_model 回退主模型。
model_name = config.skill_evolution.curator.model_name