Skip to content
Get packs

Create Data Context

Builds and maintains reusable semantic-layer skills that tell future data analyses which metrics and sources to trust.

Source Download skill
Repository
XiaomiMiMo/MiMo-Code
Stars
13,743
Licence
MIT
Files
13 files

What it does

When you ask it to save data context or create a semantic layer, the agent writes a compact, inspectable skill that records how your data should be interpreted: which metric definition is canonical, which dashboard or table is the trusted source, and what caveats to check. Later analyses can cite this layer directly. Detailed definitions and evidence go into linked reference files, and the agent points out source conflicts instead of silently picking one. It saves the result either to your personal Skills library, to a local Codex skills folder, or returns it as a portable package if no save location is available.

Great for

  • Data analysts who repeatedly answer questions about the same metrics and want consistent definitions
  • Teams using the Data Analytics plugin who want a documented, citable semantic layer
  • Data leads who want a written source inventory of what is trusted and what is missing

Skip it if

  • Ad hoc data questions or report building; those belong to the Data Analytics index skill, not this one (wrong tool for the job)
  • Sensitive datasets where the generated files might capture row-level data or credentials; the skill tries to avoid this but you should review outputs before sharing (privacy risk)

Try asking

  • “Save data context for our revenue metrics so future analyses use the canonical ARR definition”
  • “Create a semantic layer for the marketing reporting area from our dbt models and dashboards”
  • “Update the semantic layer to add a caveat that churn excludes trial accounts”
  • “Inspect and repair our existing customer-analytics semantic layer skill”
Step-by-step guide: how to use Create Data ContextInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 13 original files, the MIT licence and an attribution note.

    create-data-context.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip create-data-context.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip create-data-context.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Point the agent at durable sources such as transformation code, tests, and maintained metric docs; these outrank chat history.
  • Review the generated semantic-layer skill and its source-inventory file before relying on it for decisions.
  • Say explicitly if you want recurring refresh; otherwise the agent will only offer it and wait for approval.

What's inside

  • agents/openai.yaml
  • plugin-author-config/automation-config.md
  • plugin-author-config/source-category-config.json
  • references/automation.md
  • references/semantic-layer/connector-playbook.md
  • references/semantic-layer/skill-template.md
  • references/semantic-layer/source-intake.md
  • references/semantic-layer/weekly-polling-automation.md
  • scripts/data_analytics_preflight.py
  • scripts/record_plugin_install_suppression.py
  • scripts/validate_data_context_contract.py
  • SKILL.md
  • tests/test_state_helpers.py

Redistributed unmodified from XiaomiMiMo/MiMo-Code under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.