Detektyw danych z lupa

Detektyw danych z lupa is an agent for Claude Code from TheJacksonCode/Agent-Architecture. It costs 92 tokens per session (1,203 once invoked), scanned A, original, MIT.

An exploratory data analysis agent that profiles a dataset before modelling or decision-making. It reports column types, missing values, distributions, correlations, unusual records, data-quality issues, and visual findings.

In plain words
What is it for?
Use it with CSV, Parquet, or SQL data to create a data dictionary, charts, anomaly and outlier reports, correlation analysis, and a reproducible Jupyter notebook or script.
Why use it?
It helps teams discover what their data actually contains before trusting it. Early detection of missing, inconsistent, or unusual values can prevent misleading analysis and models.

Agent for Claude Code

Written for Claude Code: effort in frontmatter. Also seen: model in frontmatter.

Part of the agent-architecture-designer plugin — 62 commands, 60 agents shipped together

Good fit Use it with CSV, Parquet, or SQL data to create a data dictionary, charts, anomaly and outlier reports, correlation analysis, and a reproducible Jupyter notebook or script.

Compare 6 agents from other repositories ↓
Install with agentmods
npx agentmods add agents/thejacksoncode/agent-architecture/eda_analyst
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Clone the repo
git clone --depth 1 https://github.com/TheJacksonCode/Agent-Architecture

Made for: Claude Code.

Or install agent-architecture-designer, the plugin that ships this one along with the rest of its 62 commands, 60 agents.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for Detektyw danych z lupa

README.md
[![agentmods](https://agentmods.dev/badge/agents/thejacksoncode/agent-architecture/eda_analyst/github.svg)](https://agentmods.dev/agents/thejacksoncode/agent-architecture/eda_analyst)
Your own site
<a href="https://agentmods.dev/agents/thejacksoncode/agent-architecture/eda_analyst"><img src="https://agentmods.dev/badge/agents/thejacksoncode/agent-architecture/eda_analyst/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for Detektyw danych z lupa

Your own site · 80×15
<a href="https://agentmods.dev/agents/thejacksoncode/agent-architecture/eda_analyst"><img src="https://agentmods.dev/badge/agents/thejacksoncode/agent-architecture/eda_analyst.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 92 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,203 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00092 $0.01203
Opus 5.5 $0.00037 $0.00481
Sonnet 5.5 $0.00018 $0.00241
Haiku 4.5 $0.00009 $0.00120

Measured 1mo ago against content hash 3b8759131dd1, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

Detektyw danych z lupa scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

agents/eda_analyst.md · 73 lines

How it starts

The opening of the file, as written. The whole thing — 73 lines — stays where its author put it; the contents beside it link to each section on GitHub.

ROLE: Analityk EDA prowadzi eksploracyjna analize danych: profilowanie, wykrywanie anomalii, korelacje i wizualizacje. Jego misja to opisać dane tak, by zespół wiedział z czym pracuje zanim zbuduje model lub podejmie decyzje, i by ukryte pulapki wyszly na jaw wczesnie.

INPUT:

  • Surowy zbior danych (CSV, Parquet, tabela SQL)
  • Opis pochodzenia danych i kontekstu biznesowego
  • Pytanie badawcze lub hipoteza robocza do potwierdzenia
  • Słownik kolumn jeśli istnieje albo tylko metadata

OUTPUT:

  • Data dictionary z typami, brakami i rozkladami
  • Zestaw wizualizacji rozkladów i korelacji
  • Lista anomalii, outlierów i problemów jakości
  • Raport z kluczowymi obserwacjami i rekomendacjami
  • Notatnik Jupyter lub skrypt odtwarzalny dla zespołu

RESPONSIBILITIES:

  1. Profiluje dane kolumna po kolumnie z typami, brakami i unikalnościami
  2. Buduje histogramy, boxploty i density plots wychwytujące kształt rozkladu
  3. Liczy macierz korelacji Pearsona i Spearmana wykrywając kolinearności
  4. Identyfikuje wartości odstające metodami IQR, z-score i wizualnie
  5. Wykrywa paradoksy Simpsona i ukryte warunkowania zmienne
  6. Oznacza braki systematyczne (MCAR vs MAR vs MNAR) z konsekwencjami
  7. Rysuje pair ploty i cross-tabulacje dla zmiennych kategorycznych
  8. Tworzy odtwarzalny notatnik który można przekazać modelarzom i statystykowi

RULES:

  • Profilowanie danych: Opisuje każda kolumne: typ, braki, unikalne wartości, rozklad, statystyki. Tworzy data dictionary jako wspólna podstawe dla zespołu.
  • Wizualizacje rozkladów: Buduje histogramy, boxploty i density ploty dla zmiennych liczbowych oraz bar charty dla kategorycznych. Patrzy na skew, kurtoze i modalność.
  • Korelacje i relacje: Liczy macierze korelacji, pair ploty i cross-tabulacje. Szuka potencjalnych leak features i kolinearności które popsulyby model.
  • Anomalie i raport: Identyfikuje wartości odstające, braki systematyczne, duplikaty i paradoksy typu Simpson. Pisze raport z kluczowymi obserwacjami i rekomendacjami.

WHAT YOU DO NOT DO:

  • Nie buduje modeli ML ani nie trenuje sieci neuronowych
  • Nie prowadzi testów hipotez z rygorem statystycznym (to statistician)
  • Nie podejmuje decyzji biznesowych na podstawie danych
  • Nie robi ETL ani nie czyste dane poza wykryciem problemów
  • Nie projektuje schematu bazy danych (to db_architect)
  • Nie przypisuje braki metodami zaawansowanymi bez konsultacji
  • Nie wyciąga wniosków causal z danych obserwacyjnych

ANTI-PATTERNS:

  • Jump to Model - pominiecie EDA i przejście od razu do trenowania modelu
  • Mean Addiction - opisywanie zmiennych tylko średnia bez rozkladu i odchylenia
  • Correlation Equals Causation - wnioski przyczynowe z korelacji w danych obserwacyjnych
  • Outlier Removal Without Reason - wycinanie outlierów bo przeszkadzają wykresom
  • Aggregated Blindness - patrzenie tylko na agregaty globalne ignorując segmenty i Simpson paradox

REPORT FORMAT:

Summary

  • [key findings/actions taken]

Read the full file on GitHub · 73 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 1mo ago First seen · 73 lines · 92 tokens per session scan A 3b8759131dd1

Subscribe to this mod's changes

Detektyw danych z lupa is an agent published in the GitHub repository TheJacksonCode/Agent-Architecture (108 stars, last pushed 24d ago), licensed MIT. It adds 92 tokens to every session and 1,203 once invoked, about $0.0004 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-16.

Related

Other agents, from other repositories

prompt-engineer

Sharpens an existing system prompt into a tighter, more concrete, more testable one. Use when reviewing or improving a prompt rather than authoring one cold.

emtcmca/promptsmith · 36 tokens

data-engineer

Use this agent when working with data ingestion, ETL pipelines, data validation, preprocessing, schema design, or data storage. For example: building a data loading pipeline from CSV/Parquet, adding pandera schema validation, creating preprocessing transforms, setting up DVC for data versioning, optimizing data…

xvirobotics/metaskill · 76 tokens

ml-engineer

Use this agent when working with model architecture, training loops, loss functions, optimizers, hyperparameter tuning, experiment tracking, or model evaluation. For example: building a PyTorch model, writing a training loop with mixed precision, setting up an Optuna hyperparameter sweep, configuring MLflow experiment…

xvirobotics/metaskill · 81 tokens

staff-data-sci

Personas are Opus-only. The Data Science Reviewer: data science, ML, and statistics review.

dbc-oduffy/coordinator-claude · 26 tokens

data-engineer

Use this agent when working with data ingestion, ETL pipelines, data validation, preprocessing, schema design, or data storage. For example: building a data loading pipeline from CSV/Parquet, adding pandera schema validation, creating preprocessing transforms, setting up DVC for data versioning, optimizing data…

morganmuli/metaskill · 76 tokens

ml-engineer

Use this agent when working with model architecture, training loops, loss functions, optimizers, hyperparameter tuning, experiment tracking, or model evaluation. For example: building a PyTorch model, writing a training loop with mixed precision, setting up an Optuna hyperparameter sweep, configuring MLflow experiment…

morganmuli/metaskill · 81 tokens