claudient 0.1.0

package/prompts/system-prompts/nl/data-pipeline.md

@@ -0,0 +1,78 @@
+> 🇳🇱 Dit is de Nederlandse vertaling. [Engelse versie](../data-pipeline.md).
+
+# CLAUDE.md Starter — Data Pipeline Project
+
+Zet dit in de `CLAUDE.md` van je project en vul de secties tussen haakjes in.
+
+---
+
+```markdown
+# [Projectnaam] — Claude Code Instructies
+
+## Wat dit is
+[Eén alinea: welke data deze pipeline verwerkt, bronsystemen, bestemming, zakelijk doel]
+
+## Stack
+- Orchestrator: [Airflow / Prefect / Dagster / dbt Cloud]
+- Transformatie: [dbt / PySpark / Pandas / Polars]
+- Warehouse: [BigQuery / Snowflake / Redshift / DuckDB]
+- Ingestie: [Fivetran / Airbyte / aangepast]
+- Taal: [Python / SQL]
+- Infra: [Terraform op AWS / GCP / Azure]
+
+## Projectstructuur
+dbt/ (bij gebruik van dbt)
+├── models/
+│   ├── staging/      ← 1:1 met brontabellen, alleen lichte opschoning
+│   ├── intermediate/ ← Bedrijfslogica, joins
+│   └── marts/        ← Definitieve bedrijfsentiteiten (fct_-, dim_-prefixen)
+├── macros/           ← Herbruikbare SQL-macro's
+├── seeds/            ← Statische referentiedata
+└── tests/            ← Aangepaste enkelvoudige tests
+
+pipelines/ (bij gebruik van Airflow/Prefect/Dagster)
+├── dags/ / flows/    ← Pipeline-definities
+├── operators/        ← Aangepaste operators/taken
+└── utils/            ← Gedeelde hulpprogramma's
+
+## Dataconventies
+- Stagingmodellen: hernoem naar snake_case, cast typen, geen joins, geen bedrijfslogica
+- Facttabellen: fct_-prefix, één rij per gebeurtenis/transactie
+- Dimensietabellen: dim_-prefix, één rij per entiteit
+- Gebruik nooit SELECT * in productiequery's
+- Alle mart-modellen moeten unique + not_null tests hebben op primaire sleutel
+- Versheidscontroles vereist op alle bronnen
+
+## Beslissingen (niet opnieuw bespreken)
+- [Incrementele vs. volledige-verversings-strategie voor facttabellen]
+- [Tijdzone: alle tijdstempels in UTC]
+- [Granulariteit: wat één rij in elke mart-tabel vertegenwoordigt]
+- [Strategie voor laat-aankomende data]
+
+## Testvereisten
+- Elk stagingmodel: not_null op primaire sleutel
+- Elk mart-model: unique + not_null op primaire sleutel, relaties op externe sleutels
+- Bronversheid: waarschuwen bij [X] uur, fout bij [Y] uur
+
+## Prestatieregels
+- Partitioneer grote tabellen op datum — filter altijd op partitiekolom
+- Gebruik incrementele modellen voor tabellen > [X] rijen
+- Voer nooit volledige verversingen uit in productie zonder goedkeuring
+- Cluster/sorteersleutels: [specificeer bij gebruik van Snowflake/Redshift]
+
+## Commando's
+- dbt run --select staging — staginglaag uitvoeren
+- dbt test — alle tests uitvoeren
+- dbt docs generate && dbt docs serve — documentatie voorbeelden bekijken
+- dbt source freshness — versheid van brondata controleren
+
+## Nooit doen
+- Nooit bedrijfslogica in stagingmodellen plaatsen
+- Nooit datums hardcoderen — gebruik dbt-variabelen of macro's
+- Nooit echte credentials committen — gebruik omgevingsvariabelen of secrets manager
+- Nooit dbt run uitvoeren in productie zonder dat dbt test slaagt
+```
+
+---
+
+> **Werk met ons:** Claudient wordt ondersteund door [Uitbreiden](https://uitbreiden.com/) — we bouwen AI-producten en B2B-oplossingen met ontwikkelaarsgemeenschappen. [uitbreiden.com](https://uitbreiden.com/)

package/prompts/system-prompts/nl/saas-backend.md

@@ -0,0 +1,71 @@
+> 🇳🇱 Dit is de Nederlandse vertaling. [Engelse versie](../saas-backend.md).
+
+# CLAUDE.md Starter — SaaS Backend
+
+Zet dit in de `CLAUDE.md` van je project en vul de secties tussen haakjes in.
+
+---
+
+```markdown
+# [Projectnaam] — Claude Code Instructies
+
+## Wat dit is
+[Eén alinea: wat het product doet, wie het gebruikt, welk probleem het oplost]
+
+## Stack
+- Taal: [TypeScript / Python / Go]
+- Framework: [Express / FastAPI / Gin / NestJS]
+- Database: [PostgreSQL via Prisma / raw pg / SQLAlchemy]
+- Auth: [JWT met 15-min access tokens + 7-dag refresh tokens / Clerk / Auth0]
+- Cache: [Redis]
+- Queue: [BullMQ / SQS / Celery]
+- Deployment: [AWS ECS / Fly.io / Railway]
+
+## Projectstructuur
+src/
+├── api/          ← Route-handlers — slank, delegeer aan services
+├── services/     ← Bedrijfslogica — geen HTTP-concerns
+├── db/           ← Databasequeries — geen bedrijfslogica
+├── middleware/   ← Auth, rate limiting, foutafhandeling
+├── models/       ← Typedefinities en schema's
+└── utils/        ← Pure functies, geen bijeffecten
+
+## Conventies
+- Route-handlers zijn slank: invoer valideren, service aanroepen, antwoord retourneren
+- Services bevatten alle bedrijfslogica: ze weten niet van HTTP
+- DB-laag bevat alleen queries: geen bedrijfslogica, geen HTTP-concerns
+- Alle databasetoegang gaat via de db/-laag — roep ORM nooit direct aan vanuit services
+- Fouten propageren naar boven met context — slik nooit stilletjes in
+- Alle API-routes retourneren: 200 (succes), 201 (aangemaakt), 204 (geen inhoud), 400 (slechte invoer), 401 (niet auth), 403 (verboden), 404 (niet gevonden), 409 (conflict), 422 (validatie), 500 (onverwacht)
+
+## Beslissingen (niet opnieuw bespreken)
+- [Auth-mechanisme besloten: JWT, geen sessies]
+- [ORM-keuze: Prisma — geen ruwe SQL behalve voor complexe analysequeries]
+- [Foutformaat: { error: string, code: string } — wijzig de shape nooit]
+- [Geen barrel-bestanden — importeer direct vanuit bron]
+
+## Testen
+- Integratietests raken een echte testdatabase — geen DB-mocks
+- Unittests voor pure bedrijfslogica in services/
+- Testbestand: [bestandsnaam].test.ts naast het bronbestand
+- Uitvoeren: npm test
+
+## Commando's
+- npm run dev — start ontwikkelingsserver met hot reload
+- npm test — alle tests uitvoeren
+- npm run build — productiebuild
+- npm run lint — ESLint + Prettier controle
+- npm run db:migrate — uitstaande migraties uitvoeren
+- npm run db:seed — ontwikkelingsdata seeden
+
+## Nooit doen
+- Nooit bedrijfslogica in route-handlers plaatsen
+- Nooit de database direct aanroepen vanuit route-handlers
+- Nooit ruwe databasefouten retourneren aan clients
+- Nooit .env-bestanden committen
+- Nooit `any`-type gebruiken in TypeScript
+```
+
+---
+
+> **Werk met ons:** Claudient wordt ondersteund door [Uitbreiden](https://uitbreiden.com/) — we bouwen AI-producten en B2B-oplossingen met ontwikkelaarsgemeenschappen. [uitbreiden.com](https://uitbreiden.com/)

package/prompts/system-prompts/saas-backend.md

@@ -0,0 +1,69 @@
+# CLAUDE.md Starter — SaaS Backend
+
+Drop this into your project's `CLAUDE.md` and fill in the bracketed sections.
+
+---
+
+```markdown
+# [Project Name] — Claude Code Instructions
+
+## What this is
+[One paragraph: what the product does, who uses it, what problem it solves]
+
+## Stack
+- Language: [TypeScript / Python / Go]
+- Framework: [Express / FastAPI / Gin / NestJS]
+- Database: [PostgreSQL via Prisma / raw pg / SQLAlchemy]
+- Auth: [JWT with 15-min access tokens + 7-day refresh tokens / Clerk / Auth0]
+- Cache: [Redis]
+- Queue: [BullMQ / SQS / Celery]
+- Deployment: [AWS ECS / Fly.io / Railway]
+
+## Project structure
+src/
+├── api/          ← Route handlers — thin, delegate to services
+├── services/     ← Business logic — no HTTP concerns
+├── db/           ← Database queries — no business logic
+├── middleware/   ← Auth, rate limiting, error handling
+├── models/       ← Type definitions and schemas
+└── utils/        ← Pure functions, no side effects
+
+## Conventions
+- Route handlers are thin: validate input, call service, return response
+- Services contain all business logic: they do not know about HTTP
+- DB layer contains only queries: no business logic, no HTTP concerns
+- All database access goes through the db/ layer — never call ORM directly from services
+- Errors propagate upward with context — never swallow silently
+- All API routes return: 200 (success), 201 (created), 204 (no content), 400 (bad input), 401 (unauth), 403 (forbidden), 404 (not found), 409 (conflict), 422 (validation), 500 (unexpected)
+
+## Decisions (do not re-discuss)
+- [Auth mechanism decided: JWT, no sessions]
+- [ORM choice: Prisma — no raw SQL except for complex analytics queries]
+- [Error format: { error: string, code: string } — never change the shape]
+- [No barrel files — import directly from source]
+
+## Testing
+- Integration tests hit a real test database — no DB mocks
+- Unit tests for pure business logic in services/
+- Test file: [filename].test.ts next to the source file
+- Run: npm test
+
+## Commands
+- npm run dev — start development server with hot reload
+- npm test — run all tests
+- npm run build — production build
+- npm run lint — ESLint + Prettier check
+- npm run db:migrate — run pending migrations
+- npm run db:seed — seed development data
+
+## Never do
+- Never put business logic in route handlers
+- Never call the database from route handlers directly
+- Never return raw database errors to clients
+- Never commit .env files
+- Never use `any` type in TypeScript
+```
+
+---
+
+> **Work with us:** Claudient is backed by [Uitbreiden](https://uitbreiden.com/) — we build AI products and B2B solutions with developer communities. [uitbreiden.com](https://uitbreiden.com/)

package/prompts/task-specific/changelog.md

@@ -0,0 +1,81 @@
+# Prompt: Generate Changelog
+
+Use this prompt to generate a user-facing changelog from git history or a list of changes.
+
+---
+
+## From git log
+
+```
+Generate a user-facing changelog for version [X.Y.Z].
+
+Git log since last release:
+[paste: git log v[prev-version]..HEAD --oneline]
+
+Rules:
+- Write for users, not developers — explain impact, not implementation
+- Group by: New Features / Improvements / Bug Fixes / Breaking Changes
+- Omit: chore commits, dependency bumps, internal refactors with no user impact
+- Breaking changes section must be first if it exists
+- Each entry: one sentence, active voice, present tense ("Adds X", "Fixes Y", "Removes Z")
+- If a commit is unclear, ask me rather than guess
+
+Format:
+## [X.Y.Z] — [YYYY-MM-DD]
+
+### Breaking Changes
+- ...
+
+### New Features
+- ...
+
+### Improvements
+- ...
+
+### Bug Fixes
+- ...
+```
+
+---
+
+## From a list of changes
+
+```
+Generate a user-facing changelog entry for these changes:
+
+[paste bullet list of what changed]
+
+Target audience: [developers using our API / end users of the web app / etc.]
+
+Rules:
+- User impact first — "You can now X" not "We added Y"
+- Group related changes under one entry
+- One sentence per entry
+- No technical jargon unless the audience is technical
+```
+
+---
+
+## Keep a Changelog format
+
+```
+Generate a changelog entry following the Keep a Changelog format (keepachangelog.com).
+
+Changes:
+[describe or paste git log]
+
+Version: [X.Y.Z]
+Date: [YYYY-MM-DD]
+
+Use these exact section headers where applicable:
+### Added
+### Changed
+### Deprecated
+### Removed
+### Fixed
+### Security
+```
+
+---
+
+> **Work with us:** Claudient is backed by [Uitbreiden](https://uitbreiden.com/) — we build AI products and B2B solutions with developer communities. [uitbreiden.com](https://uitbreiden.com/)

package/prompts/task-specific/de/changelog.md

@@ -0,0 +1,83 @@
+> 🇩🇪 Dies ist die deutsche Übersetzung. [Englische Version](../changelog.md).
+
+# Prompt: Changelog generieren
+
+Diesen Prompt verwenden, um einen benutzerorientierten Changelog aus der Git-Historie oder einer Liste von Änderungen zu generieren.
+
+---
+
+## Aus dem Git-Log
+
+```
+Generate a user-facing changelog for version [X.Y.Z].
+
+Git log since last release:
+[paste: git log v[prev-version]..HEAD --oneline]
+
+Rules:
+- Write for users, not developers — explain impact, not implementation
+- Group by: New Features / Improvements / Bug Fixes / Breaking Changes
+- Omit: chore commits, dependency bumps, internal refactors with no user impact
+- Breaking changes section must be first if it exists
+- Each entry: one sentence, active voice, present tense ("Adds X", "Fixes Y", "Removes Z")
+- If a commit is unclear, ask me rather than guess
+
+Format:
+## [X.Y.Z] — [YYYY-MM-DD]
+
+### Breaking Changes
+- ...
+
+### New Features
+- ...
+
+### Improvements
+- ...
+
+### Bug Fixes
+- ...
+```
+
+---
+
+## Aus einer Änderungsliste
+
+```
+Generate a user-facing changelog entry for these changes:
+
+[paste bullet list of what changed]
+
+Target audience: [developers using our API / end users of the web app / etc.]
+
+Rules:
+- User impact first — "You can now X" not "We added Y"
+- Group related changes under one entry
+- One sentence per entry
+- No technical jargon unless the audience is technical
+```
+
+---
+
+## Keep a Changelog Format
+
+```
+Generate a changelog entry following the Keep a Changelog format (keepachangelog.com).
+
+Changes:
+[describe or paste git log]
+
+Version: [X.Y.Z]
+Date: [YYYY-MM-DD]
+
+Use these exact section headers where applicable:
+### Added
+### Changed
+### Deprecated
+### Removed
+### Fixed
+### Security
+```
+
+---
+
+> **Mit uns arbeiten:** Claudient wird von [Uitbreiden](https://uitbreiden.com/) unterstützt — wir bauen KI-Produkte und B2B-Lösungen mit Entwickler-Communities. [uitbreiden.com](https://uitbreiden.com/)

package/prompts/task-specific/de/debugging.md

@@ -0,0 +1,78 @@
+> 🇩🇪 Dies ist die deutsche Übersetzung. [Englische Version](../debugging.md).
+
+# Prompt: Strukturiertes Debugging
+
+Diesen Prompt verwenden, wenn ein Bug zur systematischen Diagnose an Claude übergeben wird.
+
+---
+
+## Vollständiger Debugging-Prompt
+
+```
+I have a bug. Help me diagnose it systematically.
+
+## The symptom
+[Describe exactly what goes wrong: error message, unexpected output, wrong behavior.
+Include the exact error text if there is one.]
+
+## Reproduction
+[Describe how to reproduce it — steps, curl command, test case, or "cannot reliably reproduce"]
+
+## What I've already tried
+[List anything you've already checked or changed — saves Claude from repeating it]
+
+## Relevant files
+[List the files most likely involved — or ask Claude to identify them]
+
+## Recent changes
+[What changed recently in this area? Paste relevant git log output if helpful]
+
+---
+
+Using this information:
+
+1. Identify the 3 most likely root causes, ranked by probability
+2. For each cause, describe a specific test that would confirm or rule it out
+3. Start with the highest-probability cause
+4. Do not fix anything yet — diagnose first
+
+After each test, report what you found and move to the next hypothesis or proceed to fix.
+```
+
+---
+
+## Schnellvariante (offensichtliche Bugs)
+
+Für Bugs, bei denen die Ursache ungefähr bekannt ist:
+
+```
+Bug: [describe the symptom]
+Likely cause: [your hypothesis]
+Relevant code: [paste the relevant function or file]
+
+Confirm whether my hypothesis is correct, then apply the minimal fix.
+Write a regression test after fixing.
+```
+
+---
+
+## Produktionsvorfall-Variante
+
+Für Live-Vorfälle, bei denen Schnelligkeit wichtig ist:
+
+```
+PRODUCTION INCIDENT — [service name] is [symptom]
+
+Immediate priority: identify the blast radius and any actions to reduce user impact NOW.
+
+Then diagnose:
+- Error logs: [paste relevant log lines]
+- Recent deployments: [list last 3 deploys with timestamps]
+- Affected users: [number/percentage if known]
+
+What is the fastest safe mitigation while we find the root cause?
+```
+
+---
+
+> **Mit uns arbeiten:** Claudient wird von [Uitbreiden](https://uitbreiden.com/) unterstützt — wir bauen KI-Produkte und B2B-Lösungen mit Entwickler-Communities. [uitbreiden.com](https://uitbreiden.com/)

package/prompts/task-specific/de/pr-description.md

@@ -0,0 +1,69 @@
+> 🇩🇪 Dies ist die deutsche Übersetzung. [Englische Version](../pr-description.md).
+
+# Prompt: PR-Beschreibung generieren
+
+Diesen Prompt verwenden, um eine qualitativ hochwertige Pull-Request-Beschreibung aus den eigenen Änderungen zu generieren.
+
+---
+
+## Prompt
+
+```
+Write a pull request description for the following changes.
+
+Branch: [branch name]
+Base: main
+
+Changed files:
+[paste: git diff --stat main...HEAD]
+
+What this PR does:
+[one sentence — the user should fill this in so Claude has intent, not just diff]
+
+Write the PR description with these sections:
+
+## Summary
+[2–4 bullet points: what changed and why, from the reader's perspective]
+
+## Motivation
+[one short paragraph: what problem this solves or what improvement it makes]
+
+## Changes
+[bullet list of meaningful changes — not a file list, but behavioral/structural changes]
+
+## Test plan
+[bulleted checklist of how a reviewer can verify this works:
+- specific steps to test
+- edge cases to check
+- what a successful result looks like]
+
+## Notes
+[anything a reviewer needs to know: breaking changes, migration steps, follow-up work, known limitations]
+
+Rules:
+- No generic filler ("this PR improves the codebase")
+- No file lists (the diff already shows files)
+- Specific and concrete — a reviewer should be able to test this without asking questions
+- If there are no notes, omit the Notes section
+```
+
+---
+
+## Schnellvariante (kleine PRs)
+
+Für kleine, offensichtliche Änderungen:
+
+```
+Write a concise PR description for this change: [describe the change in one sentence]
+
+Include:
+- What: one sentence
+- Why: one sentence  
+- How to test: one or two specific steps
+
+Keep it under 100 words total.
+```
+
+---
+
+> **Mit uns arbeiten:** Claudient wird von [Uitbreiden](https://uitbreiden.com/) unterstützt — wir bauen KI-Produkte und B2B-Lösungen mit Entwickler-Communities. [uitbreiden.com](https://uitbreiden.com/)

package/prompts/task-specific/debugging.md

@@ -0,0 +1,76 @@
+# Prompt: Structured Debugging
+
+Use this prompt when handing a bug to Claude for systematic diagnosis.
+
+---
+
+## Full debugging prompt
+
+```
+I have a bug. Help me diagnose it systematically.
+
+## The symptom
+[Describe exactly what goes wrong: error message, unexpected output, wrong behavior.
+Include the exact error text if there is one.]
+
+## Reproduction
+[Describe how to reproduce it — steps, curl command, test case, or "cannot reliably reproduce"]
+
+## What I've already tried
+[List anything you've already checked or changed — saves Claude from repeating it]
+
+## Relevant files
+[List the files most likely involved — or ask Claude to identify them]
+
+## Recent changes
+[What changed recently in this area? Paste relevant git log output if helpful]
+
+---
+
+Using this information:
+
+1. Identify the 3 most likely root causes, ranked by probability
+2. For each cause, describe a specific test that would confirm or rule it out
+3. Start with the highest-probability cause
+4. Do not fix anything yet — diagnose first
+
+After each test, report what you found and move to the next hypothesis or proceed to fix.
+```
+
+---
+
+## Quick variant (obvious bugs)
+
+For bugs where the cause is roughly known:
+
+```
+Bug: [describe the symptom]
+Likely cause: [your hypothesis]
+Relevant code: [paste the relevant function or file]
+
+Confirm whether my hypothesis is correct, then apply the minimal fix.
+Write a regression test after fixing.
+```
+
+---
+
+## Production incident variant
+
+For live incidents where speed matters:
+
+```
+PRODUCTION INCIDENT — [service name] is [symptom]
+
+Immediate priority: identify the blast radius and any actions to reduce user impact NOW.
+
+Then diagnose:
+- Error logs: [paste relevant log lines]
+- Recent deployments: [list last 3 deploys with timestamps]
+- Affected users: [number/percentage if known]
+
+What is the fastest safe mitigation while we find the root cause?
+```
+
+---
+
+> **Work with us:** Claudient is backed by [Uitbreiden](https://uitbreiden.com/) — we build AI products and B2B solutions with developer communities. [uitbreiden.com](https://uitbreiden.com/)

package/prompts/task-specific/es/changelog.md

@@ -0,0 +1,83 @@
+> 🇪🇸 Esta es la traducción en español. [Versión en inglés](../changelog.md).
+
+# Prompt: Generar Changelog
+
+Usa este prompt para generar un changelog orientado al usuario a partir del historial de git o una lista de cambios.
+
+---
+
+## Desde git log
+
+```
+Generate a user-facing changelog for version [X.Y.Z].
+
+Git log since last release:
+[paste: git log v[prev-version]..HEAD --oneline]
+
+Rules:
+- Write for users, not developers — explain impact, not implementation
+- Group by: New Features / Improvements / Bug Fixes / Breaking Changes
+- Omit: chore commits, dependency bumps, internal refactors with no user impact
+- Breaking changes section must be first if it exists
+- Each entry: one sentence, active voice, present tense ("Adds X", "Fixes Y", "Removes Z")
+- If a commit is unclear, ask me rather than guess
+
+Format:
+## [X.Y.Z] — [YYYY-MM-DD]
+
+### Breaking Changes
+- ...
+
+### New Features
+- ...
+
+### Improvements
+- ...
+
+### Bug Fixes
+- ...
+```
+
+---
+
+## Desde una lista de cambios
+
+```
+Generate a user-facing changelog entry for these changes:
+
+[paste bullet list of what changed]
+
+Target audience: [developers using our API / end users of the web app / etc.]
+
+Rules:
+- User impact first — "You can now X" not "We added Y"
+- Group related changes under one entry
+- One sentence per entry
+- No technical jargon unless the audience is technical
+```
+
+---
+
+## Formato Keep a Changelog
+
+```
+Generate a changelog entry following the Keep a Changelog format (keepachangelog.com).
+
+Changes:
+[describe or paste git log]
+
+Version: [X.Y.Z]
+Date: [YYYY-MM-DD]
+
+Use these exact section headers where applicable:
+### Added
+### Changed
+### Deprecated
+### Removed
+### Fixed
+### Security
+```
+
+---
+
+> **Trabaja con nosotros:** Claudient está respaldado por [Uitbreiden](https://uitbreiden.com/) — construimos productos de IA y soluciones B2B con comunidades de desarrolladores. [uitbreiden.com](https://uitbreiden.com/)