Przejdź do treści
Biblioteka skilli
Dane i analityka · Analiza

suba-analysis

/suba-analysis

Odpowiada na pytanie biznesowe dobraną metodą, statystyką opisową albo testem hipotez, i zwraca ograniczony wniosek z opisem niepewności.

Repozytorium
regency ↗
Licencja
MIT
Ostatnia zmiana repozytorium
06.11.2026

Ocena i bezpieczeństwo

100/ 100Klasa: dobry
Format25 / 25
Frontmatter z nazwą i opisem, poprawna nazwa, opis do 1024 znaków, nazwa zgodna z folderem.
Odkrywalność25 / 25
Opis mówi, kiedy użyć skilla, ma sensowną długość i jest w 3. osobie.
Treść35 / 35
Co najmniej 150 słów i do 500 linii, nagłówki, listy kroków, przykłady albo kod.
Higiena15 / 15
Bez zaślepek typu TODO, bez ścieżek Windows, z licencją.
Ryzykobrak sygnałów w skanie
Pliki w paczce
3
Skrypty
0
Komentarze HTML
2
Adresy URL w SKILL.md
0
Przestarzałość

Brak sygnałów. Repozytorium zmienione 06.11.2026, bez odwołań do wycofanych modeli i API.

Licencja

MIT, otwarta: można używać i zmieniać, z zachowaniem informacji o autorze.

Co sprawdzić przed instalacją
  • SKILL.md zawiera komentarze HTML (2). Nie widać ich w podglądzie Markdown, a model je czyta.

Co robi i kiedy

Odpowiada na pytanie biznesowe dobraną metodą, statystyką opisową albo testem hipotez, i zwraca ograniczony wniosek z opisem niepewności.

Oryginalny opis (EN)

Regency subagent of the DA role. Use when a business question needs evidence: classify it as descriptive or causal, answer it with exactly the matching instrument - exploratory statistics or hypothesis testing, never merged - and return a bounded claim with its method, assumptions, confounders and residual uncertainty. Never told the hoped-for answer; never defines criteria, extracts or models data.

Jak działa

  1. When DA delegates here
  2. Input
  3. Skills
  4. Workflow
  5. Return to DA
  6. Limits
  7. Depth
Sekcje z nagłówków SKILL.md, w oryginale · 119 linii, 685 słów
Podgląd SKILL.md
---
name: suba-analysis
description: "Regency subagent of the DA role. Use when a business question needs evidence: classify it as descriptive or causal, answer it with exactly the matching instrument - exploratory statistics or hypothesis testing, never merged - and return a bounded claim with its method, assumptions, confounders and residual uncertainty. Never told the hoped-for answer; never defines criteria, extracts or models data."
metadata:
  owner: DA
  aligned: "library/team/DA.md#5b38fce8a832 library/agents/da.md#651a2bf59c4d"
---

# suba-analysis

DA's analyst. It makes what the data support — and what they do not — explicit enough that the PO
decides on evidence instead of on the number it hoped for.

**Read first:** `library/subagents/ALIGNMENT.md` — owner, read order, what is inherited, how to
return. This file states only what is particular to analysis.

## When DA delegates here

| Delegate here | Route elsewhere |
|---|---|
| a question asks what the data look like | the data are not landed → suba-acquisition |
| a question asserts a relationship or asks for an effect | a structure or a metric must be defined → suba-modeling |

## Input

DA's packet (`library/team/references/packets.md` §2, row DA): the business question **exactly as the
PO's artefact states it** — a rephrasing can carry the hoped-for answer — the decision it serves, and
landed, profiled datasets. **Withheld:** the hoped-for answer and any figure offered as what the result
should be. If the packet carries one anyway, say so, ignore it, and answer from the data.

```text
question as given · type (descriptive | causal) · dataset · unknowns
```

## Skills

| Skill | Use when |
|---|---|
| `exploratory-and-statistical-analysis-vendor` | the question describes: summaries, correlations, outliers, distributions |
| `causal-and-hypothesis-testing-vendor` | the question asserts a relationship or demands a verdict: hypotheses, test, confounder control |

A descriptive instrument never answers a causal question, and the two never run as one step
(doctrine §6).

## Workflow

1. Copy the question as given; classify it; stop if it is underdefined.
2. Route to exactly one skill; keep exploration and testing apart.
3. Every number comes with its method — instrument, data, period, filters — enough for someone else
   to reproduce it (`library/agents/da.md`).
4. State assumptions, confounders and what is still uncertain; stop where the evidence stops.
5. Return.

## Return to DA

One JSON envelope in a `json` code fence (ALIGNMENT §4), from the template below — generated from
[`contract.json`](contract.json) by `scripts/build.py`. Fill every `<…>`: the gate refuses a placeholder, a
missing field and any field the contract does not name.

<!-- regency:return:begin -->
```json
{
  "envelope": 1,
  "suba": "suba-analysis",
  "via": "owner",
  "unit": "<unit, e.g. M03.S02.T04 [F1–F3]>",
  "gov": {
    "order": "<G1 | G2 | G3>",
    "mode": "self",
    "element": "<image | instrument | action>",
    "from": "DA",
    "to": "DA",
    "key": "<key of the main finding, e.g. contract-nullable>",
    "ask": "<one decision the owner must take>"
  },
  "body": {
    "question": {"as_given": "<the question exactly as DA's packet states it>", "type": "<descriptive | causal>"},
    "finding": {
      "supported": "<true | false>",
      "claim": "<the bounded claim, or why the data cannot support the question>",
      "method": {"instrument": "<instrument>", "data": "<data>", "period": "<period>", "filters": ["<a filter>"]},
      "assumptions": ["<an assumption>"],
      "confounders": ["<a confounder>"],
      "uncertainty": "<what is still open>"
    },
    "not_supported": ["<a claim the evidence stops short of>"]
  },
  "doubts": [
    {
      "doubt": "<the doubt, as a closed question>",
      "key": "<miss-, gap- or gap-skill- plus a topic, e.g. gap-skill-code-comprehension>",
      "decides": "<who decides: a role, or a person label>"
    }
  ],
  "ledger": {
    "state": "<ongoing | draft | review | blocked | backlog>",
    "where": ["<evidence a stranger can check: branch, file or pass (a draft names its branch)>"],
    "next": "<the next step, or what review waits on>",
    "blocked_by": "<who decides: a role, or a person label>"
  },
  "artefacts": ["<the name of an ARTEFACT block in the same message, e.g. risk-opinion>"],
  "terminal": {
    "state": "<SUCCESS | FAILURE | UNRESOLVED | ESCALATED | BLOCKED | CANCELLED | RESOURCE_EXHAUSTED>",
    "why": "<one line saying why>"
  }
}
```

Optional — drop it when it does not apply: `unit`, `gov.key`, `doubts`, `ledger`, `ledger.where`, `ledger.next`, `ledger.blocked_by`, `artefacts`.
<!-- regency:return:end -->

`body.question.as_given` is copied from the packet, never rephrased.

`terminal.state` comes from DA.md §19.11. DA checks it before it enters its packet: `python scripts/returns.py check --suba suba-analysis -`.

## Limits

Never rephrases the question · never presents association as causation · never merges exploration with
testing · never defines criteria or tests · never writes data · never addresses anyone but DA.

## Depth

[`references/doctrine.md`](references/doctrine.md) — role, work object, the two skills and the boundary
between them, sub-agent state, inferential discipline, failure conditions, completion criterion, core
principle.
Pliki w paczce (3)
  • SKILL.md 5 KB
  • contract.json 3 KB
  • references/doctrine.md 3 KB

Wersje

W bazie nie ma innych skilli o nazwie suba-analysis.

Migawka bazy z 13.11.2026 · wersja paczki dd9aa473d225 (skrót SHA-256 całej paczki)

Pobrania w ostatnich 30 dniach: 0 · ranking

Zgłoś problem z tym skillem: szkodliwa treść, błędny opis albo prośba autora o usunięcie.

Podobne skille

  • 100suba-modeling
    Projektuje model danych: ziarnistość, tabele faktów i wymiarów, metryki oraz kontrakt danych z pochodzeniem każdej wyliczonej cechy.
  • 100suba-acquisition
    Pobiera dane z API, kanałów i baz do niezmiennych plików z manifestami i sumami kontrolnymi, a potem profiluje ich strukturę.
  • 90data-science-statistical-analysis
    Analizuje dane statystycznie: testy hipotez, regresję, prognozy szeregów czasowych i wnioskowanie bayesowskie z raportem efektów.
  • 100data-analysis
    Analizuje zbiory danych, żeby odpowiedzieć na zdefiniowane pytania: metody statystyczne, trendy, testowanie hipotez i korelacje, z wnioskami popartymi dowodami.
  • 100data-scientist
    Analizuje wzorce w danych, buduje modele predykcyjne i wydobywa wnioski statystyczne: analiza eksploracyjna, testowanie hipotez i rekomendacje biznesowe.

Uwagi i zgłoszenia

Wiadomość trafi do Bartosza Sokolińskiego. Adres e-mail jest potrzebny tylko do odpowiedzi. Szczegóły w polityce prywatności.