# Investigation > Scaffolds a template for a small investigation for empirical experimentation. Use when the user asks for an investigation, experiment or research to support making design decisions, assess performance or try out a new idea that requires data or an MVP. Use when this capability is needed. - Skill: `tomevault-io/investigation-2` (Agent Skill, multi-file: 2 files) - Install (CLI): `npx skillmds@latest add tomevault-io/investigation-2` - Raw SKILL.md: https://api.skillmd.com/api/skills/tomevault-io/investigation-2/raw - Safety review: pending (external: skill-scanner PASS, skillspector PASS) - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Research & Search - Author: tomevault-io (https://skillmd.com/u/tomevault-io) - Updated: 2026-09-17 - Page: https://skillmd.com/skills/tomevault-io/investigation-2 --- # Set up an investigation ## Instructions 1. Create a folder in `{REPO_ROOT}/scratch/` with the format `{YYYY-MM-DD}-{descriptive-name}`. You will mainly work inside this folder. 2. Create a `README.md` in this folder where you will: 1. Describe the task. 2. Take background notes for context. 3. Create a task list for progress tracking. 4. Summarize results and outcomes. 3. Create scripts and testing data files in this folder to conduct an empirical investigation. 4. For multi-step investigations with sub-experiments, consider generating individual markdown files to take interediate notes. 5. The `README.md` should contain actionable findings to support a subsequent implementation or design decision. ## Best practices - Both for this library and other ones, it's worth writing a simple script or calling the library interactively to list its members or try out different constructions to explore the API and document it in a markdown file called `API.md`. - Generate figures when applicable (e.g., plots, data visualizations) and reference them inline in the markdown files. - Always use `uv` with self-contained dependencies when using Python. - When you need a local web server, use `npx serve -p 8080 --cors --no-clipboard &` (run in background to avoid blocking). - Prefer self-contained HTML+JS pages (without React if possible) and use Playwright MCP for rendering screenshots. - Use subagents whenever possible to save context and delegate tasks in parallel. ## Important: Scratch work is gitignored The `scratch/` directory is in `.gitignore` and will NOT be committed. When distilling findings into PRs: - Do NOT assume local scratch notes will be checked in - Include all relevant information inline in PR descriptions or code comments - Copy key findings, code snippets, benchmark results, or data directly into the PR - The PR should be self-contained and not rely on scratch files for context --- > Converted and distributed by [TomeVault](https://tomevault.io/claim/talmolab) — claim your Tome and manage your conversions.