← ClaudeAtlas

wiki_asklisted

Chat with the knowledge base. Answers user questions by dynamically employing RAG context extraction, graph database queries, and targeted keyword searches to provide cited, hallucination-free answers.
Misaka16384/Wikify · ★ 2 · Data & Documents · score 68
Install: claude install-skill Misaka16384/Wikify
# LLM Wiki — Knowledge Base Q&A Skill (wiki_ask) > **Resolving script paths (read first):** Commands below invoke scripts as `<BIN>/X.py` (and a few as `<SKILLS>/...`). Resolve these to **absolute paths once** before running anything: > > - `<SKILL_DIR>` = the directory this `SKILL.md` lives in. > - `<SKILLS>` = the `skills/` folder containing this skill = `<SKILL_DIR>/..` > - `<BIN>` = the `bin/` folder beside it = `<SKILL_DIR>/../../bin` > > Do **not** hardcode a fixed prefix like `.agents/bin` or `../bin`: shell relative paths resolve against the current working directory (usually the topic root), not this skill's location. Once resolved, `<BIN>` is typically `.agents/bin` when invoked from the hub root, or `.claude/bin` from inside a topic directory. This skill transforms the agent into a conversational interface for the knowledge base. When the user asks a question, you must intelligently deploy a combination of deterministic search tools to retrieve context and provide a highly rigorous, cited answer. > **Tooling (framework-agnostic):** This skill is written tool-agnostic. Where it says *file-read tool*, use your agent's equivalent (`Read` in Claude Code, `view_file` in Antigravity, `cat`/open elsewhere). Shell commands run via `Bash`/`PowerShell` (Claude Code) or your framework's shell tool. ## Core Rule: Zero Hallucination > **CRITICAL INSTRUCTION**: Do NOT answer questions using your parametric memory. Every factual claim or summary you provide MUST be grounded i