tech
/tech
Sprawdza i przygotowuje profile technologii do Rawkode Academy: wiersze w D1, terminy do transkrypcji i logotypy w R2.
- Autor
- rawkode-academy ↗
- Repozytorium
- rawkode-academy ↗
- Licencja
- MIT
- Ostatnia zmiana repozytorium
- 27.10.2026
Ocena i bezpieczeństwo
- Pliki w paczce
- 2
- Skrypty
- 0
- Komentarze HTML
- 0
- Adresy URL w SKILL.md
- 9
Brak sygnałów. Repozytorium zmienione 27.10.2026, bez odwołań do wycofanych modeli i API.
MIT, otwarta: można używać i zmieniać, z zachowaniem informacji o autorze.
Co robi i kiedy
Sprawdza i przygotowuje profile technologii do Rawkode Academy: wiersze w D1, terminy do transkrypcji i logotypy w R2.
Oryginalny opis (EN)
Research, validate, and set up Rawkode Academy technology profiles: platform-technologies D1 rows, transcription terms, and R2 logos. Use when adding a technology, checking an existing technology profile, or updating its metadata and content.
Jak działa
- 1. Research & Discovery
- 2. Check Existence
- 3. The Row
- 4. Rich Text
- 5. Write
- 6. Logos
- 7. Invalidate the Website Cache
- 8. Validate & Update (if it exists)
- Guardrails
Podgląd SKILL.md
---
name: tech
description: "Research, validate, and set up Rawkode Academy technology profiles: platform-technologies D1 rows, transcription terms, and R2 logos. Use when adding a technology, checking an existing technology profile, or updating its metadata and content."
---
# Rawkode Academy Technology Profiles
Technology profiles are rows in the `platform-technologies` D1 database
(`projects/rawkode.academy/platform/technologies`), served through the GraphQL
gateway (`https://api.rawkode.academy`) to the website. Logos live in the R2
bucket `rawkode-academy-content`. The repository has no technology files any
more: never create or edit Markdown/MDX for a technology.
When the user asks to add or check a technology (e.g. "Add Bun", "Check
Kubernetes"), follow this process.
## 1. Research & Discovery
Research only from primary project sources: the official site, documentation,
source repository and official licensing information. Gather:
- **Official name** and a concise, objective description of what it is.
- **Official website** and **documentation** URLs.
- **Source repository** (GitHub/GitLab) and **license** (SPDX id, e.g. `Apache-2.0`, `MIT`).
- **Logos**: official vector logos (SVG) for `icon`, `horizontal` and `stacked`, with provenance (the URL you took each from).
- **Key features**, **use cases**, **core concepts** and **ecosystem** (related tools or integrations).
- **Spellings** people say or mistype (the name, product names, CLI names): these become transcription terms.
Prefer primary sources and distinguish verified facts from uncertain claims.
## 2. Check Existence
The id is the kebab-case public slug (`kubernetes`, `bun`). Check the live API
first, then the database:
```bash
curl -fsS https://api.rawkode.academy/ -H 'content-type: application/json' \
--data '{"query":"query($id:String!){ technologyById(id:$id){ id name website documentation source license status category subcategory logos { icon horizontal stacked } } }","variables":{"id":"<id>"}}'
cd projects/rawkode.academy/platform/technologies
cuenv exec -e production -- bunx wrangler d1 execute platform-technologies --remote \
--config read-model/wrangler.jsonc \
--command "SELECT id, name, website, documentation, source, license, status, category, subcategory, logos, updated_at FROM technologies WHERE id = '<id>' OR name = '<Name>'"
```
Read nearby technologies (same category) the same way for prose and metadata
conventions. The allowed values for `status`, `category`, `subcategory` and
`matrix` live in `projects/rawkode.academy/website/src/lib/explorer/dimension-values.ts`;
the column shapes in `projects/rawkode.academy/platform/technologies/data-model/schema.ts`.
## 3. The Row
Table `technologies` (D1 `platform-technologies`). JSON columns are stored as
JSON text.
| Column | Notes |
|---|---|
| `id` | kebab-case slug, primary key |
| `name` | official name, unique |
| `description` | the profile body in **Markdown** (NOT NULL; kept until every reader uses `body`) |
| `body` | the same profile as a TipTap/ProseMirror JSON document (see 4) |
| `website`, `documentation`, `source` | URLs |
| `license` | SPDX id, default `Unknown` |
| `status` | `alpha`, `beta`, `stable`, `preview`, `superseded`, `deprecated` or `abandoned` (default `stable`) |
| `category`, `subcategory` | one singular value each |
| `aliases`, `related_technologies`, `use_cases`, `features` | JSON string arrays (`related_technologies` holds technology ids, no `/index`) |
| `learning_resources` | JSON `{ "official"?: [], "community"?: [], "tutorials"?: [] }` |
| `logos` | JSON `{ "icon"?: true, "horizontal"?: true, "stacked"?: true }`: set a flag only once that file is in R2 |
| `cncf`, `community`, `seo` | JSON objects (keys in `data-model/schema.ts`) |
| `matrix` | JSON; required `status` (`skip`, `watch`, `explore`, `learn`, `adopt`, `advocate`, `graveyard`, `guilty-pleasure`), optional `grouping` (`plumbing`/`platform`/`observability`/`security`), `confidence` (`gut`/`some-experience`/`deep-experience`), `trajectory` (`rising`/`stable`/`falling`) |
| `updated_at` | milliseconds: `CAST(strftime('%s','now') AS INTEGER) * 1000` |
Add optional fields only when verified and useful. Do not add obsolete fields
(`categories`, `radar`).
Transcription spellings are rows in `platform-transcription-terms`, table
`transcription_terms` (`kind` = `'technology'`, `foreignId` = the technology id,
`term`).
## 4. Rich Text
`body` is a TipTap document as written by `@rawkodeacademy/rich-text`. Write the
profile as Markdown (concise, objective, grounded in the sources; for a new
profile, sections such as *What is X?*, *Core concepts*, *Getting started*,
*Common use cases*, *Ecosystem*), then convert and validate it:
```bash
cd packages/rich-text
bun -e 'import { fromMarkdown } from "./src/markdown/index.ts"; import { validateDoc } from "./src/schema.ts"; const file = process.argv[1]; const doc = fromMarkdown(await Bun.file(file).text(), { mdx: false, file }); validateDoc(doc); console.log(JSON.stringify(doc));' /absolute/path/to/body.md > /absolute/path/to/body.json
```
Store the Markdown in `description` and the JSON in `body`. Never hand-write
the JSON; a document `validateDoc` rejects must not be stored.
## 5. Write
Write the SQL to a file outside the repo (single quotes doubled inside string
literals, JSON as text), review it, then apply it:
```sql
INSERT INTO technologies (id, name, description, body, website, documentation, source, license, status, category, subcategory, logos, updated_at)
VALUES ('bun', 'Bun', '<markdown>', '<body json>', 'https://bun.sh', 'https://bun.sh/docs', 'https://github.com/oven-sh/bun', 'MIT', 'stable', '<category>', '<subcategory>', '{"icon":true}', CAST(strftime('%s','now') AS INTEGER) * 1000);
```
For an update, `UPDATE technologies SET <only the corrected columns>, updated_at = CAST(strftime('%s','now') AS INTEGER) * 1000 WHERE id = '<id>';`
(for one JSON key: `logos = json_set(coalesce(logos, '{}'), '$.stacked', json('true'))`).
```bash
cd projects/rawkode.academy/platform/technologies
cuenv exec -e production -- bunx wrangler d1 execute platform-technologies --remote --config read-model/wrangler.jsonc --file=/absolute/path/to/technology.sql
cd ../transcription-terms
cuenv exec -e production -- bunx wrangler d1 execute platform-transcription-terms --remote --config read-model/wrangler.jsonc \
--command "INSERT OR IGNORE INTO transcription_terms (kind, foreignId, term) VALUES ('technology', '<id>', 'Bun'), ('technology', '<id>', 'bunx')"
```
## 6. Logos
Upload verified SVGs to R2 before setting their `logos` flags. Keys are
`technologies/<id>/icon.svg`, `technologies/<id>/horizontal.svg` and
`technologies/<id>/stacked.svg` (served at `https://content.rawkode.academy/<key>`):
```bash
cuenv exec -e production -- bunx wrangler r2 object put rawkode-academy-content/technologies/<id>/icon.svg \
--file=/absolute/path/to/icon.svg --content-type=image/svg+xml --remote
curl -fsS -I https://content.rawkode.academy/technologies/<id>/icon.svg
```
Never invent logos, download them without provenance or upload placeholders.
If no official asset exists, leave the flag unset and tell the user which
variants are missing.
## 7. Invalidate the Website Cache
After every D1 write or logo upload:
```bash
curl -fsS -X POST https://rawkode.academy/api/cache/invalidate \
-H "Authorization: Bearer $CACHE_INVALIDATION_TOKEN" \
-H 'content-type: application/json' \
-d '{"tags":["technologies:<id>"]}'
```
(`just invalidate-content '{"tags":["technologies:<id>"]}'` in
`projects/rawkode.academy/website` does the same.) The endpoint takes a while
to answer (it purges twice, ~50s apart). Without the token the change shows up
after the website's hourly change detection; say so.
## 8. Validate & Update (if it exists)
1. Read the row (step 2) and the rendered page `https://rawkode.academy/technology/<id>`.
2. Compare `website`, `documentation`, `source` and `license` with your research; check the profile for outdated versions, years or claims.
3. Make the narrowest correction that resolves the request: `UPDATE` only the columns that change. When changing the profile text, update `description` and regenerate `body` from it together.
4. Invalidate the cache and re-read the API to confirm the change.
## Guardrails
- Tone: informative, technical and objective.
- Do not invent metadata, URLs, capabilities, relationships or asset provenance.
- Keep changes scoped to the requested technology; no broad rewrites or unrelated rows.
- Never write `content/**` files; the website no longer reads them.
- Report what was verified, changed (rows, R2 keys), invalidated and left unresolved.
Pliki w paczce (2)
- SKILL.md 9 KB
- agents/openai.yaml 1 KB
Wersje
| Repozytorium | Wynik | Licencja |
|---|---|---|
| Ta strona · rawkode-academy/rawkode-academy | 100 | MIT |
| karloscodes/fusionaly-oss | 100 | MIT |
| tammai/rspec | 86 | MIT |
Identyczna treść występuje jeszcze w 1 innych folderach na GitHubie. Tu pokazujemy jedną, najlepiej ocenioną kopię.
Migawka bazy z 13.11.2026 · wersja paczki 0a8388b965a4 (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
- 100rawkode-academy-video-importer
Importuje filmy z YouTube do Rawkode Academy: uruchamia importer, wgrywa zasoby R2, tworzy wiersze D1, czyści cache, weryfikuje transkodowanie i transkrypcję. - 100tech-assessment
Ocenia technologie, stos, wykonalność techniczną, trendy i decyzje build vs buy oraz przygotowuje roadmapę architektury. - 100tech-advisor
Doradza w wyborze technologii na podstawie aktualnych dowodów, także oceniając jej dojrzałość.