Data Table Analysis
NVIDIA-AI-Blueprints/deep-researcher-agent
A skill your agent uses for converting researched facts or user-provided data into structured tables by writing code, then running Python/pandas calculations in the job-scoped sandbox.
Loads, cleans, aggregates and joins tabular data with pandas under a verification rule: every number reported must be one that the code actually printed.
Install Verified Data Analysis with pandas
$ npx skills add pipeshub-ai/pipeshub-ai --skill data-analysis -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install pipeshub-ai/pipeshub-ai data-analysis --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/pipeshub-ai/pipeshub-ai.git skills-src && mkdir -p .claude/skills && cp -r skills-src/backend/python/app/agents/agent_loop/skills/builtin_packs/data-analysis .claude/skills/data-analysis && rm -rf skills-srcUse ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.
Claude Code skills documentation · loads skills from .claude/skills/
Install the "data-analysis" agent skill from https://github.com/pipeshub-ai/pipeshub-ai/tree/main/backend/python/app/agents/agent_loop/skills/builtin_packs/data-analysis into .claude/skills/data-analysis/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "data-analysis", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/pipeshub-ai/pipeshub-ai/tree/main/backend/python/app/agents/agent_loop/skills/builtin_packs/data-analysisType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add pipeshub-ai/pipeshub-ai --skill data-analysis -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install pipeshub-ai/pipeshub-ai data-analysis --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/pipeshub-ai/pipeshub-ai.git skills-src && mkdir -p .agents/skills && cp -r skills-src/backend/python/app/agents/agent_loop/skills/builtin_packs/data-analysis .agents/skills/data-analysis && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "data-analysis" agent skill from https://github.com/pipeshub-ai/pipeshub-ai/tree/main/backend/python/app/agents/agent_loop/skills/builtin_packs/data-analysis into .agents/skills/data-analysis/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "data-analysis", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add pipeshub-ai/pipeshub-ai --skill data-analysis -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install pipeshub-ai/pipeshub-ai data-analysis --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/pipeshub-ai/pipeshub-ai.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/backend/python/app/agents/agent_loop/skills/builtin_packs/data-analysis .cursor/skills/data-analysis && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "data-analysis" agent skill from https://github.com/pipeshub-ai/pipeshub-ai/tree/main/backend/python/app/agents/agent_loop/skills/builtin_packs/data-analysis into .cursor/skills/data-analysis/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "data-analysis", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/pipeshub-ai/pipeshub-ai.git --path backend/python/app/agents/agent_loop/skills/builtin_packs/data-analysis--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add pipeshub-ai/pipeshub-ai --skill data-analysis -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install pipeshub-ai/pipeshub-ai data-analysis --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/pipeshub-ai/pipeshub-ai.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/backend/python/app/agents/agent_loop/skills/builtin_packs/data-analysis .gemini/skills/data-analysis && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "data-analysis" agent skill from https://github.com/pipeshub-ai/pipeshub-ai/tree/main/backend/python/app/agents/agent_loop/skills/builtin_packs/data-analysis into .gemini/skills/data-analysis/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "data-analysis", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install pipeshub-ai/pipeshub-ai data-analysisInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add pipeshub-ai/pipeshub-ai --skill data-analysis -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/pipeshub-ai/pipeshub-ai.git skills-src && mkdir -p .github/skills && cp -r skills-src/backend/python/app/agents/agent_loop/skills/builtin_packs/data-analysis .github/skills/data-analysis && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "data-analysis" agent skill from https://github.com/pipeshub-ai/pipeshub-ai/tree/main/backend/python/app/agents/agent_loop/skills/builtin_packs/data-analysis into .github/skills/data-analysis/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "data-analysis", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add pipeshub-ai/pipeshub-ai --skill data-analysis -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install pipeshub-ai/pipeshub-ai data-analysis --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/pipeshub-ai/pipeshub-ai.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/backend/python/app/agents/agent_loop/skills/builtin_packs/data-analysis .opencode/skills/data-analysis && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "data-analysis" agent skill from https://github.com/pipeshub-ai/pipeshub-ai/tree/main/backend/python/app/agents/agent_loop/skills/builtin_packs/data-analysis into .opencode/skills/data-analysis/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "data-analysis", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
Facts
- Skill name
data-analysis- GitHub stars
- 3.8k
- Token cost
- ~1.2k tokens
- SKILL.md length
- 584 words
- Files
- 1
- Skills in repo
- 9
- Repo updated
- First seen
- Licence
- Apache-2.0
At a glance
Loads, cleans, aggregates and joins tabular data with pandas under a verification rule: every number reported must be one that the code actually printed.
- Works in 3 steps: Always run df.info() and df.describe()… → Cross-check any headline number by… → Never report a number your code didn't…
- Computing totals or averages from a CSV, Excel or JSON export
- SKILL.md covers Loading, Cleaning, Aggregation, joins, pivots and Verification discipline (do…
- Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md
What it does
The skill assumes pandas, numpy and scipy are already installed in the sandbox. Its loading rules pin column types when the schema is known so ID-like values keep leading zeros, parse date columns explicitly so sorting and range filters are chronological, and name the sheet when reading Excel files instead of trusting the default first sheet. Cleaning rules ask for null and duplicate counts before any aggregation, with an explicit decision to drop, fill or flag, explicit numeric coercion for columns that only look numeric, and a before and after row count after every step that removes or changes rows.
For aggregation, each column gets its own named function in `groupby` calls instead of a bare sum over the whole frame. The guiding discipline is that reported figures come only from code output and never from inference or estimation, which suits questions such as totals by region, cleaning a dataset or joining two files on a customer ID.
When your agent uses it
- Computing totals or averages from a CSV, Excel or JSON export
- Cleaning a messy dataset without silently dropping rows
- Joining two files on a shared key and checking the row counts
Example prompts
- “What is the total revenue by region in sales_2025.csv?”
- “Join orders.xlsx and customers.csv on customer_id and tell me how many orders have no match.”
- “Clean up this survey export and report how many rows each step removes.”
Requirements
- A Python sandbox with pandas, numpy and scipy
Workflow steps
3 steps, taken from the first numbered list in SKILL.md.
- Always run df.info() and df.describe() (or .value_counts() for categorical columns) before analyzing — this catches wrong dtypes…
- Cross-check any headline number by computing it a second, independent way — e.g. if you report a total via groupby().sum(), also…
- Never report a number your code didn't actually print. If you're describing a trend or a comparison, make sure the specific figures you…
What it can do on your machine
Read from SKILL.md and the folder at commit f5aee03. It shows what the files ask for, not the result of running them.
Tool permissions
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
Runs code
No scripts in the folder and no shell commands in SKILL.md.
From the folder's file list and the shell code blocks in SKILL.md.
Network
No URLs in SKILL.md.
From URLs in SKILL.md, links to its own repository left out.
Credentials
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Context cost
Verified Data Analysis with pandas loads about 1.2k tokens when it runs. Until then it costs about 103 tokens; SKILL.md has 584 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
Safety
Auto-check passed
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.
SKILL.md
The full file from pipeshub-ai/pipeshub-ai at commit f5aee03, republished under its Apache-2.0 licence (© pipeshub-ai). 584 words, ~1,202 tokens.
Download SKILL.mdSave it as
.claude/skills/data-analysis/SKILL.md (or your agent's skills folder).- name
- data-analysis
- description
- Use this skill whenever the user asks you to load, clean, aggregate, join, or otherwise analyze tabular data (csv, xlsx, json, or a database export) — 'what's the total revenue by region', 'clean up this dataset', 'join these two files on customer id'. Establishes a verification discipline so reported numbers are always numbers the code actually printed, never numbers you inferred or estimated.
Data analysis with pandas
pandas, numpy, and scipy are already installed in the sandbox — no install_packages call needed for any workflow below.
Loading
- Pin dtypes on load rather than letting pandas infer them, when you know the schema —
pd.read_csv(path, dtype={"customer_id": str, "amount": float}). Inference is usually right but silently wrong in a specific, dangerous way for ID-like columns: a numeric-looking ID (e.g."00123") inferred asint64loses its leading zeros. - Parse date columns explicitly (
parse_dates=["order_date"]orpd.to_datetime(df["col"])post-load) rather than leaving them as strings — string dates sort and filter lexicographically, not chronologically, which silently produces wrong "most recent N" or date-range results. - For Excel sources with multiple sheets, always pass
sheet_name=explicitly (or inspectpd.ExcelFile(path).sheet_namesfirst) —read_excel's default of the first sheet is a common source of "the numbers don't match" bugs when the data you actually need is on a different sheet.
Cleaning
- Check for nulls (
df.isnull().sum()) and duplicates (df.duplicated().sum()) before any aggregation — decide explicitly whether to drop, fill, or flag them, rather than letting them silently skew asum/mean(pandas aggregations skip NaN by default, which is not always the right call — a "total" that silently excludes rows with a missing value is a different number than "total of all rows", and the user should know which one they're getting). - Coerce types explicitly after cleaning (
pd.to_numeric(df["col"], errors="coerce")) rather than assuming a column that "looks numeric" already is — mixed-type columns from a manually-maintained spreadsheet are common, and a silent string comparison instead of a numeric one produces wrong sort/filter results without raising any error. - After any cleaning step that drops or modifies rows, print the before/after row count — a cleaning step that unexpectedly drops 80% of the data is a bug, not a result, and should be caught immediately rather than discovered after the whole analysis is built on the reduced dataset.
Show full SKILL.md (281 more words)Show less
Aggregation, joins, pivots
groupby(...).agg(...)for aggregates; always specify the aggregation function explicitly per column (.agg({"amount": "sum", "order_id": "count"})) rather than a bare.sum()across the whole frame, which silently includes every numeric column whether or not it makes sense to sum it.- For joins (
pd.merge), always specifyhow=explicitly (don't rely on the"inner"default when the user's intent might be"left"/"outer") and check the resulting row count against expectations — an unexpected row-count change after a merge (either fewer rows than the left frame, or way more) almost always means a many-to-many relationship you didn't account for, or a key-matching problem (mismatched types, whitespace, or casing in the join column). pivot_tablefor cross-tabulations; passfill_value=0explicitly when a missing combination should read as zero rather thanNaN, since which one is correct depends on the question being asked.
Verification discipline (do this before reporting any number)
- Always run
df.info()anddf.describe()(or.value_counts()for categorical columns) before analyzing — this catches wrong dtypes, unexpected nulls, and outliers early, before they propagate into a wrong conclusion. - Cross-check any headline number by computing it a second, independent way — e.g. if you report a total via
groupby().sum(), also sanity-check it againstdf["amount"].sum()on the ungrouped frame (should match), or against a manual filter-and-sum on a known subset. Two different code paths agreeing is meaningfully more trustworthy than one path you didn't double-check. - Never report a number your code didn't actually print. If you're describing a trend or a comparison, make sure the specific figures you state in your response are ones that appeared in your program's printed output — do not round, estimate, or extrapolate a number that wasn't literally computed and displayed.
© pipeshub-ai, Apache-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
Files
Just SKILL.md in backend/python/app/agents/agent_loop/skills/builtin_packs/data-analysis of pipeshub-ai/pipeshub-ai.
Open the folder on GitHubat commit f5aee03
Compare with similar skills
Verified Data Analysis with pandas next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Verified Data Analysis with pandas this skillpipeshub-ai/pipeshub-ai | 3.8k | — | ~1.2k | Automated safety check: Pass | Apache-2.0 | |
| Data Table AnalysisNVIDIA-AI-Blueprints/deep-researcher-agent | 883 | — | ~2.5k | Automated safety check: Pass | Apache-2.0 | |
| Data AnalysisEXboys/skilllite | 170 | — | ~176 | Automated safety check: Pass | MIT | |
| CSV and Excel MergerOneWave-AI/claude-skills | 323 | — | ~1.6k | Automated safety check: Pass | MIT | |
| Codebookbrycewang-stanford/Auto-Empirical-Research-Skills | 4.5k | — | ~527 | Automated safety check: Notes | Custom licence | |
| Python Executorcortega26/chile-hub | 113 | 2 repos | ~1.5k | Automated safety check: Pass | MIT |
Similar skills
Data Analysis
EXboys/skilllite
Analyze CSV/JSON data with statistics, filtering, and aggregation.
CSV and Excel Merger
OneWave-AI/claude-skills
Combines CSV, TSV and Excel files into one verified table with pandas, by stacking or joining, mapping columns, normalizing keys and removing duplicates.
Codebook
brycewang-stanford/Auto-Empirical-Research-Skills
Auto-generates a Markdown codebook from a dataset (CSV, DTA, Excel, Parquet) with types and summary statistics.
Python Executor
cortega26/chile-hub
Execute Python code in a safe sandboxed environment via [inference.sh](https://inference.sh).
Pandas Pro
Jeffallan/claude-skills
Handles pandas DataFrame work: cleaning, merging, groupby aggregation, pivots, time-series resampling and memory tuning, with checks on dtypes, shapes and nulls.
More from pipeshub-ai/pipeshub-ai
All 9 skills in this repoPipesHub Excel Spreadsheet Builder
pipeshub-ai/pipeshub-ai
Creates and edits .xlsx workbooks with real Excel formulas rather than hardcoded computed values, defaulting to exceljs in TypeScript with a static formula-safety check.
Office Open XML Utilities
pipeshub-ai/pipeshub-ai
Unpacks a .docx or .pptx into pretty-printed XML, lets you make small targeted edits, and repacks it into a file Office will open.
PDF Generation, Forms and Extraction
pipeshub-ai/pipeshub-ai
Picks the right library for generating a new PDF, filling an existing PDF form, or extracting text and tables, defaulting to Node where possible.
PowerPoint Deck Builder
pipeshub-ai/pipeshub-ai
Creates new PowerPoint decks with pptxgenjs in TypeScript, reads existing decks with python-pptx, and applies a design-quality checklist so every slide has real visual hierarchy.
Chart Type Selection Guide
pipeshub-ai/pipeshub-ai
Picks the right chart type for a data question and applies readability rules like axis labels, colorblind palettes and legend restraint.
Word Document Creation And Editing
pipeshub-ai/pipeshub-ai
Routes a Word document request to the right approach: a TypeScript library for new files, XML editing for existing ones, and plain reading only.
Related
Works with
Categories
Questions about Verified Data Analysis with pandas
What does Verified Data Analysis with pandas do?
Loads, cleans, aggregates and joins tabular data with pandas under a verification rule: every number reported must be one that the code actually printed. The skill assumes pandas, numpy and scipy are already installed in the sandbox. Its loading rules pin column types when the schema is known so ID-like values keep leading zeros, parse date columns explicitly so sorting and range filters are chronological, and name the sheet when reading Excel files instead of trusting the default first sheet.
When should I use Verified Data Analysis with pandas?
Verified Data Analysis with pandas fits situations like: computing totals or averages from a CSV, Excel or JSON export; cleaning a messy dataset without silently dropping rows; joining two files on a shared key and checking the row counts.
How do I install Verified Data Analysis with pandas in Claude Code?
Run `npx skills add pipeshub-ai/pipeshub-ai --skill data-analysis -a claude-code`. Or copy the skill folder (backend/python/app/agents/agent_loop/skills/builtin_packs/data-analysis in pipeshub-ai/pipeshub-ai) into .claude/skills/data-analysis in your project. Claude Code loads it when a task matches its description.
How do I install Verified Data Analysis with pandas in Codex?
Run `npx skills add pipeshub-ai/pipeshub-ai --skill data-analysis -a codex`. Or copy the skill folder (backend/python/app/agents/agent_loop/skills/builtin_packs/data-analysis in pipeshub-ai/pipeshub-ai) into .agents/skills/data-analysis in your project. Codex loads it when a task matches its description.
Can I use Verified Data Analysis with pandas in Cursor, Gemini CLI or GitHub Copilot?
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add pipeshub-ai/pipeshub-ai --skill data-analysis -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/data-analysis, .gemini/skills/data-analysis, .github/skills/data-analysis and .opencode/skills/data-analysis in your project.
What does Verified Data Analysis with pandas need to run?
SKILL.md names no scripts, command-line tools or credentials: Verified Data Analysis with pandas is instructions for the agent only. Our summary lists: A Python sandbox with pandas, numpy and scipy.
Does Verified Data Analysis with pandas access the network?
SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.
Is Verified Data Analysis with pandas safe to install?
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.
What licence does Verified Data Analysis with pandas use?
Verified Data Analysis with pandas is published under the Apache-2.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.
How many tokens does Verified Data Analysis with pandas use?
About 1.2k tokens (SKILL.md is roughly 4.8k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.
What are the alternatives to Verified Data Analysis with pandas?
Skills that share tags, products or a category with Verified Data Analysis with pandas: Data Table Analysis (NVIDIA-AI-Blueprints/deep-researcher-agent, 883 stars), Data Analysis (EXboys/skilllite, 170 stars), CSV and Excel Merger (OneWave-AI/claude-skills, 323 stars) and Codebook (brycewang-stanford/Auto-Empirical-Research-Skills, 4.5k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
Who maintains Verified Data Analysis with pandas?
pipeshub-ai (a GitHub organization) maintains it in pipeshub-ai/pipeshub-ai, which has 3,816 GitHub stars. The repository holds 9 skills in this directory. The repository was last updated on October 8, 2026.
Source: pipeshub-ai/pipeshub-ai on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.