Zeabur rag inspect

Skill zeabur/rag-service/skills/zeabur-rag-inspect

View a single chunk's full content plus related reports, search signals, and edit history. Use before editing, when investigating why a result appeared, or when reviewing a reported issue. Requires admin scope.From its SKILL.md

Install

npx -y skills add zeabur/rag-service --skill zeabur-rag-inspect

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

5 stars5 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

2.6 KB, as published. Nobody here has run it

RAG — Inspect

Get the full picture of a single knowledge base chunk: its content, related reports, search signals that surfaced it, and audit history.

Base URL: $ZEABUR_RAG_URL Auth: Authorization: Bearer $RAG_API_KEY — admin scope required

API

curl -s "$ZEABUR_RAG_URL/api/admin/chunks/<chunk_id>" \
  -H "Authorization: Bearer $RAG_API_KEY"

Input: Chunk ID (e.g. DOC-example, LEARNED-1712345678-abc123).

Response

{
  "chunk": {
    "id": "DOC-example",
    "title": "...",
    "question": "...",
    "answer": "...",
    "text_content": "...",
    "tags": ["..."],
    "source": "docs",
    "verified": true,
    "status": "verified",
    "parent_id": null,
    "created_at": "...",
    "url": "..."
  },
  "reports": [
    { "id": "...", "type": "outdated", "query": "...", "detail": "...", "status": "open", "created_at": "..." }
  ],
  "signals": [
    { "id": "...", "query": "...", "mode": "hybrid", "top_similarity": 0.012, "feedback_score": null, "created_at": "..." }
  ],
  "audit_log": [
    { "id": "...", "chunk_id": "DOC-example", "action": "edit", "old_value": {}, "new_value": {}, "created_at": "..." }
  ]
}

Key fields

text_content — the text used for embedding and BM25 index. You MUST read this before editing a chunk, because zeabur-rag-edit requires you to provide updated text_content whenever you change title/question/answer.
reports — any open or closed reports pointing to this chunk. If there's an open report, it may need action.
signals — recent queries that surfaced this chunk in search results. Shows how the chunk is being found and whether users found it helpful (via feedback_score).
audit_log — edit history. Check this to see who changed what and when.

Common workflows

Before editing: inspect → read text_content → construct updated text_content in the same format → call zeabur-rag-edit.
Investigating a report: inspect the reported chunk → check if the report is valid → decide to edit, reject, or close the report.
Understanding search behavior: inspect a chunk that keeps appearing in irrelevant queries → check signals to see which queries surface it → consider editing text_content to improve relevance.

Error

Returns 404 with {"error": "Chunk not found"} if the ID doesn't exist.