csv
Read, parse, write, and clean CSV files — dialect handling, type inference, and validation.
SKILL.mdskills/documents/csv/SKILL.md
GitHubCSV
Purpose
Work with CSV files: parse with the right dialect, validate, clean, and write.
When to Use
- User shares a CSV
- User wants to export data as CSV
- User has a messy CSV and wants it cleaned
Source
Custom skill, written for this library.
Notes
Pairs with data-analysis.
When NOT to Use
Do not use when the request is outside this capability, required context is unavailable, or a safer specialized skill applies.
Capabilities
- Apply the csv workflow consistently.
- Produce a clear, reviewable result.
- Surface assumptions, constraints, and unresolved risks.
Inputs
- The user's goal and desired output.
- Relevant source material, constraints, and environment details.
- Acceptance criteria when available.
Workflow
- Confirm the goal, scope, and constraints.
- Inspect the available context before acting.
- Apply the skill-specific guidance in this document.
- Verify the result and report limitations or next steps.
Tools
- No mandatory tool unless declared in frontmatter.
- Use only project-approved tools and documented optional dependencies.
Examples
Requests that should activate this skill include: "read this CSV"; "parse CSV"; "clean this CSV".
Safety
- Preserve user data and existing work.
- Confirm before destructive or externally visible actions.
- Do not expose credentials or claim unverified results.