메뉴
Create a new knowledge base within the system
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
| name | kb-init |
| description | Create a new knowledge base within the system |
| user_invocable | true |
| arguments | Optional: KB name. If omitted, Claude will ask interactively. |
Create and configure a new knowledge base.
Gather info — If $ARGUMENTS provides a name, use it. Otherwise, ask the user:
machine-learning, zk-security). Will become the directory name under kbs/.Validate — Read kbs.yaml and check:
Create the KB — Three steps:
a. Add entry to kbs.yaml:
<name>:
path: kbs/<name>
description: "<description>"
# private: true (only if user requested)
b. Create the directory:
mkdir -p kbs/<name>
c. Rebuild the index to register the new KB:
python3 .kb/kb-index.py build --kb <name>
Confirm — Tell the user:
kbs/<name>//kb-question --kb <name> "your question"/kb-search --kb <name> "query"/kb-question --kb <name>?"kbs.yamldefault: true unless the user explicitly asks — there should only be one default KB$ARGUMENTS
Find near-duplicate notes and merge them safely. Use when the user asks to "dedupe the KB", "find duplicate notes", "merge these notes", or after large research sessions that may have re-covered existing ground.
Re-verify time-sensitive claims in KB notes against live sources and apply visible corrections. Use when the user asks to "refresh the KB", "re-verify notes", "check what's gone stale", or on a schedule. Distinct from /kb-review (structural quality) — this checks whether the facts are still true.
Synthesize a coherent narrative explanation of a topic from multiple KB notes. Use when the user asks to "explain X from the KB", wants a structured overview/briefing woven from existing notes, or asks "what's the full picture on X".
Answer a single question and save the answer as an atomic note in the knowledge base. Use when the user poses a discrete factual/conceptual question they want captured for later, says "look this up and save it", "add this to the KB", or "note down the answer to X".
Agentic research loop that deeply investigates a topic via web search and saves findings as atomic notes
Audit the knowledge base for quality issues (orphans, broken links, missing citations, stale notes, tag drift) and suggest improvements. Use when the user asks to "review the KB", "check note quality", "find orphans/broken links/duplicates", or "what needs cleanup".