wiki_researchlisted

# LLM Wiki — Research Skill (wiki_research) > **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 deep, parallel academic research, spinning up multi-perspective subagents to drill into complex topics and compile unified verdicts. > **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), *web-search* (`WebSearch` in Claude Code, `search_web`), *shell* (`Bash`/`PowerShell`). Use the closest equivalent your framework provides; if a parallel sub-agent tool is unavailable, investigate each research dimension sequentially yourself. When the user asks to research a topic: 1. **Draft a Dynamic Research Plan**: Analyze the user's research query and determine the domain (e.g., Mathematics, Theoretical Physics, Computer Science). Subdivide the query into 3 or more distinct, domain-specific investigative dimensions. * *Example (Physics/Math)*: "Axiomatic Consistency Auditor", "Phenomenology & Experimental Reviewer", "Theoretical Extrapolator". * *Example (CS)*: "Technical Deep Dive", "Critical Reviewer", "Empirical Auditor". * *