wiki_enrichlisted

# LLM Wiki — Enrich Skill (wiki_enrich) > **CLI (read first):** This skill drives the `magi` CLI (MAGI research workspace tool, assumed installed on PATH). If unsure of your surroundings, run `magi sync` first to locate the workspace. For the full syntax of any command: `magi <command> --help`. This skill handles "gap-filling" for concept extraction. Due to token limits, initial compilations might miss secondary theorems, lemmas, or physics corollaries. This skill ensures high concept density. > **Tooling (framework-agnostic):** This skill is written tool-agnostic. Map each capability to your own agent's tool — *read-file* (`Read` in Claude Code, `view_file` in Antigravity), *sub-agent / parallel task* (`Task`/`Agent` in Claude Code, `invoke_subagent` in Antigravity), *shell* (`Bash`/`PowerShell`). Use the closest equivalent your framework provides; if a parallel sub-agent tool is unavailable, mine the chunks sequentially yourself. When the user asks to enrich, deep-dive, or查漏补缺 (fill gaps) on a paper: 0. **Pre-enrichment Placeholder Fill (Main Agent)**: * Run the placeholder detection command: `magi wiki placeholders "<compiled_file>"` * For each detected placeholder (e.g. `[STUB: Awaiting synthesis]`), scan the original `raw/papers/` source file to see if an "exactly matching semantic context" exists. * **Anti-Hallucination Guardrail**: If you find what you believe to be the matching context, **prompt the user for confirmation** before surgica