atlan-author-knowledge
/atlan-author-knowledge
Tworzy i edytuje pakiety wiedzy w Atlan, np. glosariusz, opisy tabel i runbooki, które czytają agenty.
- Autor
- atlanai ↗
- Repozytorium
- atlan-plugins ↗
- Licencja
- Apache-2.0
- Ostatnia zmiana repozytorium
- 12.11.2026
Ocena i bezpieczeństwo
- Pliki w paczce
- 6
- Skrypty
- 0
- Komentarze HTML
- 0
- Adresy URL w SKILL.md
- 0
Brak sygnałów. Repozytorium zmienione 12.11.2026, bez odwołań do wycofanych modeli i API.
Apache-2.0, otwarta: można używać i zmieniać, z zachowaniem licencji i informacji o zmianach.
Co robi i kiedy
Tworzy i edytuje pakiety wiedzy w Atlan, np. glosariusz, opisy tabel i runbooki, które czytają agenty.
Oryginalny opis (EN)
Create or edit a knowledge bundle in Atlan, the curated Markdown agents read, such as a glossary, table docs or runbooks. Use when turning notes into Atlan knowledge or changing a term. Not for metric definitions (analytics-build).
Jak działa
SKILL.md bez nagłówków drugiego poziomu · 37 linii, 449 słów
Podgląd SKILL.md
---
name: atlan-author-knowledge
description: Create or edit a knowledge bundle in Atlan, the curated Markdown agents read, such as a glossary, table docs or runbooks. Use when turning notes into Atlan knowledge or changing a term. Not for metric definitions (analytics-build).
license: Apache-2.0
---
# Author knowledge
In search mode, the default, run each tool named here through the runner for what it does
(`mcp__call_read_tool` for a read, `mcp__call_write_tool` for a create or update,
`mcp__call_tool` for anything destructive), with `session_id` and a first-person `rationale`
beside `name` and `arguments`; your first call opens
the session by sending `objective`. Once the session is open, a `registry__` read this host also lists by name can be called directly, with `session_id` and `rationale` as its own arguments and no runner. Send independent reads
together in one turn, or as one `mcp__call_read_tools` call (up to 8) when the host lists it. A create names its `workspace_id`: the one the user names, else
`.atlan/share.json`'s at the project root, else theirs from `registry__list_workspaces` with
`{"kind": "user_default"}`. Write first, and read `registry__get_kind_schema` only after a
refusal that names no field. Open `skills/atlan-get-started/references/calling-tools.md` only for
a failed call, a limit, a source key or full mode. Replies lead with the result, in four sentences or fewer and under 80 words (a table or a list the user asked for aside). Check each: no greeting, apology or closing offer; counts rather than lists, handles only when asked; one next step, not a menu; nothing about how tools work or calls skipped or retried. Ask before every write.
A read that fails with 502, 503, 504 or `retryable: true` is sent again unchanged, 3 tries
in all (waiting `retry_after_seconds` first when the error gives it), before you tell the user
Atlan is not answering. Never repeat a write: read it back first.
Before the first write, check that this organization has the knowledge app:
`mcp__describe_tools { "names": ["extension__knowledge__create"] }`. When it is not listed, or a
call answers `unknown_tool`, say in one line that the organization does not have Atlan's
knowledge app and an admin can install it. Do not record the bundle as a generic `file` or
artifact in its place.
A metric's definition of record is the `metric` and its saved SQL that the Analytics add-on's
`analytics-build` publishes, not a bundle. A bundle may explain a metric in prose for agents, with
`resource` pointing at the published metric, but never restates its calculation; when the user
wants to define or change a metric, use `analytics-build`, or say the Analytics add-on does it.
Follow [the knowledge workflow](references/author.md) for creation, reading and editing.
Load only the references that procedure calls for.
Read pinned bundle bytes before editing. Knowledge edits use the app's `apply_changes`
workflow with `base_version_ordinal`; preserve its conflict handling. On a stale base,
re-read and reconcile the user's changes before retrying. Do not substitute a generic
artifact append or overwrite another editor's work. Validate the complete bundle and
read back the resulting version.
Pliki w paczce (6)
- SKILL.md 3 KB
- agents/openai.yaml 1 KB
- references/author.md 7 KB
- references/bundle-format.md 3 KB
- references/read-a-bundle.md 2 KB
- references/update-a-bundle.md 3 KB
Wersje
W bazie nie ma innych skilli o nazwie atlan-author-knowledge.
Migawka bazy z 13.11.2026 · wersja paczki 2cf94ef399c8 (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
- 86atlan-author-agents
Rejestruje lub aktualizuje agenta w Atlanie: łączy go z repozytorium i jego skillami, rejestruje narzędzie i przygotowuje dane logowania. - 100atlan-find-artifacts
Przeszukuje rejestr Atlan pod kątem skilli, agentów, wiedzy i wyników oraz właścicieli i użytkowników. - 100atlan-search
Pomaga znaleźć dane w katalogu Atlan: który z narzędzi MCP (semantic_search, search_assets, get_assets, traverse_lineage) użyć i jakie filtry dają właściwy wynik przy pytaniach o tabele, dashboardy i tagi. - 100accelerator-readiness
Sprawdza gotowość laptopa do akceleratora analityki agentowej: agentów kodujących, połączenie z hurtownią i narzędzia BI, potem pisze raport. - 95okf-authoring
Tworzy paczki wiedzy w formacie Open Knowledge Format z plików markdown z nagłówkami YAML na podstawie źródeł danych. - 100okf
Tworzy, aktualizuje i wykorzystuje paczki wiedzy Open Knowledge Format (markdown z YAML) opisujące usługi, API, schematy, metryki, runbooki i decyzje.