AgentSkillsCN

nova-comments

管理NovaDB评论——创建、读取、更新、删除并搜索对象上的评论。

SKILL.md
--- frontmatter
name: nova-comments
description: Manage NovaDB comments — create, read, update, delete, and search comments on objects.

Body Format: XHTML with <div> Root

Comment bodies must be valid XHTML wrapped in <div>. Plain text and other root elements (<p>, <span>) cause HTTP 400.

code
"<div>Missing Email</div>"
"<div><p>Paragraph one</p><p>Paragraph two</p></div>"

Tools

OperationTool
Createnovadb_cms_create_comment(branchId, objectRef, body)
Readnovadb_cms_get_comment(commentId)
Listnovadb_cms_get_comments(branchRef?, objectRef?, user?)
Updatenovadb_cms_update_comment(commentId, body)
Deletenovadb_cms_delete_comment(commentId)
Searchnovadb_index_search_comments(branch, filter)
Countnovadb_index_count_comments(branch, filter)
Facetsnovadb_index_comment_occurrences(branch, ...)

Create

code
novadb_cms_create_comment(
  branchId=<branchId>,       # numeric
  objectRef=<objectId>,      # numeric
  body="<div>Text</div>",
  username="john.doe"        # optional, defaults to $cms
)

Search (Index API)

code
novadb_index_search_comments(
  branch="<branchId>",
  filter={ "searchPhrase": "...", "user": "...", "mentioned": "...", "objectTypes": [...] }
)

sortField: 0=Id, 1=User, 2=Created, 3=Modified, 4=ObjectId, 5=ObjectType. Pagination via skip/take.

Gotchas

  • •<div> root is mandatory — other root elements or plain text cause HTTP 400
  • •branchId and objectRef are numeric — not apiIdentifier or GUID
  • •commentId is a string
  • •CMS pagination uses take + continue token; Index pagination uses skip + take