text
| 1 | # Dataset Profiler |
| 2 | |
| 3 | Most wrong analyses are not wrong in the statistics, they are wrong in the assumptions |
| 4 | made about the data in the first ten minutes. This skill front-loads that: what each |
| 5 | column actually contains, where the nulls are and whether they are random, which rows |
| 6 | are duplicates, and which questions this dataset simply cannot answer. |
| 7 | |
| 8 | ## When to use |
| 9 | |
| 10 | - Any dataset you did not create yourself. |
| 11 | - Before the first chart, model, or aggregate. |
| 12 | - When a number looks surprising and you need to know whether it is real. |
| 13 | |
| 14 | ## Install |
| 15 | |
| 16 | ```bash |
| 17 | npx openagents-cli add openagents/data-cleaning |
| 18 | ``` |
| 19 | |
| 20 | | Runtime | Installed to | |
| 21 | |---|---| |
| 22 | | `claude-code` | `.claude/skills/data-cleaning/` | |
| 23 | | `codex` | `.codex/skills/data-cleaning/` | |
| 24 | | `openai-agents` | `agents/data-cleaning/` | |
| 25 | | `generic` | `.openagents/data-cleaning/` | |
| 26 | |
| 27 | ## What is in the package |
| 28 | |
| 29 | - `SKILL.md` - the profiling procedure and what to report. |
| 30 | - `checks.md` - the specific checks per column type, with what each failure implies. |
| 31 | |
| 32 | ## License |
| 33 | |
| 34 | MIT |
| 35 |