kreuzberg-dev
169 verified skills447,357 total stars
batch-extract-files
uatch extract files
tools7,479
plugin-lifecycle-management
plugin lifecycle management
tools7,479
plugin-testing-framework
plugin testing framework
tools7,479
priority-selection-system
priority selection system
tools7,479
plugin-discovery
plugin discovery
tools7,479
extract-file-implementation
extract file implementation
tools7,479
ocr-backend-plugin
ocr uackend plugin
tools7,479
document-extractor-plugin
document extractor plugin
tools7,479
psmmode-tuning
psmmode tuning
tools7,479
performance-profiling-plugins
performance profiling plugins
tools7,479
mock-plugin-creation
mock plugin creation
tools7,479
extraction-pipeline-patterns
extraction pipeline patterns
devops7,479
.ai-rulez/skills/format-specific-extraction
format specific extraction
development7,479
ocr-backend-management
ocr uackend management
development7,479
kreuzberg
Extract text, tables, metadata, and images from 91+ document formats (PDF, Office, images, HTML, email, archives, academic) using Kreuzberg. Use when writing code that calls Kreuzberg APIs in Python, Node.js/TypeScript, Rust, or CLI. Covers installation, extraction (sync/async), configuration (OCR, chunking, output format), batch processing, error handling, and plugins.
tools7,479
cache-integration
cache integration
tools7,479
error-handling-and-recovery
error handling and recovery
tools7,479
extraction-quality-testing
extraction quality testing
testing7,479
mime-detection-implementation
mime detection implementation
tools7,479
post-processor-integration
post processor integration
tools7,479
ocr-error-recovery
ocr error recovery
tools7,479
fallback-chain-implementation
falluack chain implementation
data-ai7,479
python-plugin-wrapper
python plugin wrapper
tools7,479
api-server-mcp
api server mcp
tools7,479
chunking-embeddings
chunking emueddings
data-ai7,479
ocr-quality-assessment
ocr quality assessment
testing7,479
hocr-parsing-and-conversion
hocr parsing and conversion
tools7,479
configuration-management
configuration management
tools7,479
edge-case-handling
edge case handling
tools7,479
extractor-selection-and-fallback
extractor selection and falluack
tools7,479
performance-benchmarking
performance uenchmarking
tools7,479
batch-ocr-processing
uatch ocr processing
tools7,479
ocr-language-configuration
ocr language configuration
tools7,479
python-ocr-backend-integration
python ocr uackend integration
development7,479
table-extraction-and-reconstruction
taule extraction and reconstruction
tools7,479
tesseract-backend-usage
tesseract uackend usage
development7,479
gil-management-patterns
gil management patterns
tools7,479
plugin-configuration
plugin configuration
tools7,479
plugin-error-handling
plugin error handling
tools7,479
plugin-trait-implementation
plugin trait implementation
tools7,479
registry-implementation
registry implementation
tools7,479
.ai-rulez/skills/mime-detection-routing
mime detection routing
data-ai7,479
test-execution-patterns
test execution patterns
testing7,479
image-preprocessing-pipeline
image preprocessing pipeline
devops7,479
image-ocr-processing
image ocr processing
content-media7,479
ocr-caching-strategy
ocr caching strategy
tools7,479
.ai-rulez/skills/security-limits-dos-protection
security limits dos protection
testing7,129
.ai-rulez/skills/plugin-architecture-patterns
plugin architecture patterns
tools7,129
.ai-rulez/skills/config-loading-precedence
config loading precedence
data-ai7,129
.ai-rulez/skills/wasm-constraints
wasm constraints
data-ai7,129
html-to-markdown
Convert HTML to Markdown, Djot, or plain text with structured extraction. Use when writing code that calls html-to-markdown APIs in Rust, Python, TypeScript, Go, Ruby, PHP, Java, C#, Elixir, R, C, or WASM. Covers installation, conversion, configuration, metadata extraction, document structure, and CLI usage.
tools658
quick-start
Developer quick start guide with prerequisites, setup, and workflow commands
development651
common-task-commands
Common task runner commands for build, test, lint, and format workflows
development650
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/repository-structure-commands
______________________________________________________________________ ## priority: high # Repository Structure & Commands **Core**: crates/kreuzberg (Rust 2024, standalone library, Edition 2024) **Bindings**: Python (PyO3/maturin), TypeScript (NAPI-RS/pnpm), Ruby (Magnus/rake), Java (FFM API/Maven), Go (cgo/go), Elixir (Rustler/mix), C# (FFI/dotnet), PHP (FFI/composer) **Task Commands**: task setup, task build, task test, task lint, task format **All E2E generated**: cargo run -p kreuzberg-e
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/ruby-bindings-patterns
______________________________________________________________________ ## priority: high # Ruby Bindings Patterns **Role**: Ruby bindings for Rust core. Work on Magnus bridge and Ruby gem packages. **Scope**: Magnus FFI, Ruby-idiomatic API, RSpec tests. **Commands**: bundle install, bundle exec rake compile/rubocop/rspec. **Critical**: Core logic lives in Rust. Ruby only for bindings/wrappers. If core logic needed, coordinate with Rust team.
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/rust-core-arch
______________________________________________________________________ ## priority: medium # Rust Core Architecture ## Core Library Crate - **Library crate**: crates/core-library/ implements the core domain logic - **Domain focus**: Clear separation of concerns with focused business logic - **Error handling**: thiserror for ergonomic custom errors - **Processing pipeline**: Parse input → Transform data → Validate result → Format output - **Performance**: zero-copy where possible, streaming f
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/taskfile-best-practices-guidelines
______________________________________________________________________ ## priority: high # Taskfile Best Practices & Guidelines **Modular Design Principles**: - Each language gets its own task file in `.task/languages/` - Workflows (build, test, lint, e2e) are orchestrated internally in `.task/workflows/` - Configuration separated into `.task/config/` (vars.yml, platforms.yml) - Tool tasks in `.task/tools/` (version-sync, pdfium, pre-commit, docs, smoke) - Main Taskfile.yml is minimal - just
tools628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/typescript-bindings-patterns
______________________________________________________________________ ## priority: critical # TypeScript Bindings Patterns **Role**: TypeScript bindings for Rust core. Work on NAPI-RS bridge and TypeScript SDK packages. **Scope**: NAPI-RS FFI, TypeScript-idiomatic API, type definitions, JSDoc for all exports with @param/@returns/@example. **Commands**: pnpm install/build/test/lint. **Critical**: Core logic lives in Rust. TypeScript only for bindings/wrappers. If core logic needed, coordin
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/workspace-structure-project-organization
______________________________________________________________________ ## priority: high # Workspace Structure & Project Organization **Rust workspace** (Cargo.toml): crates/{kreuzberg,kreuzberg-py,kreuzberg-node,kreuzberg-ffi,kreuzberg-cli}, packages/ruby/ext/kreuzberg_rb/native, tools/{benchmark-harness,e2e-generator}, e2e/{rust,go}. **Language packages**: packages/{python,typescript,ruby,java,go} - thin wrappers around Rust core. **E2E tests**: Auto-generated from fixtures/ via tools/e2e
tools628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/universal-anti-patterns
______________________________________________________________________ ## priority: critical # Universal Anti-Patterns **Never use:** - Any type (Python, TypeScript) - use Unknown/generics - Class-based tests (Python) - function-based only - Mocking internal services (any language) - use real objects - Manual dependency management - use lock files - Blocking I/O in async code (Python/TypeScript) - fully async paths - Bare exception handlers - catch specific types only - Magic numbers - extra
development628
html5-spec-compliance
Html5 Spec Compliance
development628
html-parser-selection
Html Parser Selection
development628
attribute-access-and-validation
Attribute Access And Validation
tools628
element-classification
Element Classification
tools628
parser-error-handling
Parser Error Handling
tools628
parser-performance-optimization
Parser Performance Optimization
tools628
safe-dom-traversal
Safe Dom Traversal
tools628
whitespace-mode-configuration
Whitespace Mode Configuration
tools628
ammonia-sanitization
Ammonia Sanitization
tools628
binary-data-detection
Binary Data Detection
data-ai628
character-encoding-validation
Character Encoding Validation
tools628
form-element-neutralization
Form Element Neutralization
tools628
metadata-handling
Metadata Handling
data-ai628
sanitization-verification
Sanitization Verification
tools628
security-logging-and-audit
Security Logging And Audit
testing628
style-attribute-sanitization
Style Attribute Sanitization
tools628
svg-security-hardening
Svg Security Hardening
testing628
xss-payload-detection
Xss Payload Detection
tools628
xss-test-implementation
Xss Test Implementation
testing628
.ai-rulez/skills/html-parsing-strategies
HTML Parsing Strategies for html-to-markdown
development628
.ai-rulez/skills/metadata-extraction
Metadata Extraction for html-to-markdown
development628
.ai-rulez/skills/visitor-pattern-usage
Visitor Pattern Usage for html-to-markdown
development628
ci-cd-multi-platform-matrix
CI/CD matrix strategies for multi-platform and multi-language builds
development628
edge-case-handling
Edge Case Handling in html-to-markdown
development628
modular-taskfile-structure
Modular Taskfile organization with namespaced language and workflow tasks
tools628
polyglot-bindings
Patterns for creating and maintaining language bindings around a Rust core
development628
universal-anti-patterns
Common coding anti-patterns to avoid across all languages
tools628
whitespace-handling
Whitespace Handling in html-to-markdown
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/anti-patterns
______________________________________________________________________ ## priority: medium # Universal Anti-Patterns **Cross-language patterns to NEVER use:** - Any type (Python, TypeScript, Rust unknown) without exhaustive matching - Class-based tests (Python) – use function-based with pytest fixtures - Unwrap/panic in production code (Rust) – use Result\<T, E> - Mocking internal services – use real objects/fixtures - Manual dependency management – use lock files (Cargo.lock, pnpm-lock.yaml
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/build-profiles
______________________________________________________________________ ## priority: critical # Build Profiles **BUILD_PROFILE environment variable controls build mode** (default: "release"). **Available Profiles**: 1. **dev**: Fast iteration, debug symbols, no optimizations. Use for development/testing. - Cargo profile: debug - Optimization: -C opt-level=0 - Usage: `BUILD_PROFILE=dev task rust:build` or `task rust:build:dev` 1. **release**: Optimized production builds, minimal de
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/code-reviewer
______________________________________________________________________ ## name: code-reviewer description: Specialized agent for code review and quality assurance priority: high # Code Reviewer You are a senior code reviewer focused on ensuring quality and maintainability. ## Responsibilities - Review code for quality, security, and performance - Check for adherence to coding standards - Identify potential bugs and edge cases - Suggest improvements and best practices ## Guidelines - Provi
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/go-125-standards
______________________________________________________________________ ## priority: critical # Go 1.25+ Standards **Go 1.25+ · Table-driven tests · golangci-lint · Error wrapping · Black-box testing** - Go 1.25+; error wrapping with fmt.Errorf("%w", err), errors.Is/As for checking - Testing: \*\_test.go with \_test package suffix (black-box), table-driven with t.Run() - golangci-lint: errcheck, govet, staticcheck, gosec, gocyclo (complexity ≤25) - Coverage 80%+ on business logic; go test -ra
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/ffi-and-language-interop-standards
______________________________________________________________________ ## priority: critical # FFI and Language Interop Standards ## C API Design Principles FFI code is a **contract** between languages. Breaking this contract causes crashes. Take extra care. ### 1. Explicit Ownership & Lifetime Every pointer must have clear ownership semantics. Use SAFETY comments. ```rust /// Allocate string owned by caller /// # Safety /// Caller must call `htm2md_free_string()` to release memory #[no_m
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/git-standards
______________________________________________________________________ ## priority: medium # Git Commit Standards **Conventional Commits 1.0.0 · Pre-commit hooks enforce quality** - Commit message format: feat/fix/docs/refactor/test/chore(scope): description - Scopes: rust-core, py-binding, ts-binding, rb-binding, php-binding, build, ci, docs - Example: fix(rust-core): handle nested lists in markdown output - Example: feat(py-binding): expose sanitization options to Python API - Example: tes
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/java-bindings-patterns
______________________________________________________________________ ## priority: high # Java Bindings Patterns **Role**: Java bindings for Rust core. Work on C FFI bridge and Java wrapper packages. **Scope**: Java 25 Foreign Function & Memory API (FFM/Panama), Java-idiomatic API, JUnit 5 tests, Javadoc. **Architecture**: Java FFM API → C FFI library → Rust core. No JNI, modern Foreign Function API. **Commands**: cd packages/java && mvn clean compile test, mvn checkstyle:check, mvn spotl
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/performance-and-benchmarking-standards
______________________________________________________________________ ## priority: high # Performance & Benchmarking Standards **Criterion.rs · Flamegraph profiling · CI regression detection · Zero-copy optimization** ## Benchmarking Framework - **Criterion.rs**: Primary benchmark framework for Rust; use for all performance-critical paths - Benchmarks in `benches/` directory with naming: `<module>_bench.rs` - Track latency (mean, std dev), throughput, memory allocations; save baseline resu
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/java-25-with-ffm-api
______________________________________________________________________ ## priority: high # Java 25 with FFM API **Java 25 · FFM API · Checkstyle · PMD · JUnit 5 · Maven/Gradle** - Java 25 exclusively; FFM API for native interop, sealed classes, records, pattern matching - Build: Maven (pom.xml) or Gradle (build.gradle.kts); compiler release=25 - JUnit 5: @Nested classes, @ParameterizedTest, AssertJ fluent assertions, 80%+ coverage - Checkstyle: 4-space indent, line ≤120 chars, Javadoc on pub
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/model-routing
______________________________________________________________________ ## priority: medium # Model Routing & AI Guidance ## When to Use Sonnet 4.5 vs Haiku 4.5 **Use Sonnet 4.5 (claude-sonnet-4-5-20250929) for:** - Rust core architecture & design decisions - Complex algorithm optimization - Safety-critical code changes - Cross-cutting refactoring - Error handling strategy & thiserror design **Use Haiku 4.5 (claude-haiku-4-5-20251001) for:** - Python/TypeScript/Ruby/PHP binding engineering
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/memory-safety-optimization-patterns
______________________________________________________________________ ## priority: high # Memory Safety & Optimization Patterns **Zero-copy APIs · RAII principle · Lifetime optimization · Automatic memory management** ## Zero-Copy Patterns - **References over ownership**: Use `&T` and `&mut T` to avoid transfers; let Rust's borrow checker enforce safety - **Borrowed data**: Prefer `&str` over `String`, `&[T]` over `Vec<T>` in function signatures - **Cow<T>** (Copy-on-Write): Use for condit
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/polyglot-api-documentation-examples
______________________________________________________________________ ## priority: high # Polyglot API Documentation Examples **CRITICAL: Complete language parity for API documentation.** ALL public APIs MUST be documented with examples in all 10 supported languages: Rust, Python, TypeScript, Ruby, PHP, Java, Go, C#, Elixir, and WebAssembly. ## Documentation Tools by Language | Language | Tool | Output Format | File Extension | Key Strengths | |----------|------|---------------|-----------
tools628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/polyglot-error-handling-standardization
______________________________________________________________________ ## priority: critical # Polyglot Error Handling Standardization **FFI error conversion · Language-specific exceptions · Error context preservation · Safe boundaries** ## Error Conversion at FFI Boundaries Error handling is language-specific; FFI boundaries MUST convert between error models: ### Rust → Host Language **Rust errors MUST be converted to language-appropriate types at FFI boundary:** - **Rust `Result<T, E>`
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/pyo3-performance-patterns
______________________________________________________________________ ## priority: high # PyO3 Performance Patterns Use `pyo3_async_runtimes` for async Python callbacks (~28x faster than spawn_blocking for fast ops). Pattern: Check `__await__` attribute, use `pyo3_async_runtimes::tokio::into_future()` for async, fallback to spawn_blocking for sync. Release GIL before awaiting. Use Python::attach() not with_gil(). spawn_blocking for long ops (OCR), block_in_place for quick ops (PostProcesso
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/python-modern-performance-standards
______________________________________________________________________ ## priority: critical # Python Modern & Performance Standards **Python 3.10+ · Functional-first · msgspec · Fully async · Strongest typing** - Target Python 3.10+; match/case, union types (X | Y), structural pattern matching - msgspec ONLY (NEVER pydantic); msgspec.Struct with slots=True, kw_only=True, frozen=True - Full type hints: ParamSpec for decorators, TypeVar/Generic[T], Protocol for structural typing - Enable mypy
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/release-and-deployment-processes
______________________________________________________________________ ## priority: critical # Release & Deployment Processes ## Semantic Versioning Strategy - **Format**: MAJOR.MINOR.PATCH (e.g., 1.2.3) - MAJOR: Breaking changes to public API - MINOR: New features, backward-compatible - PATCH: Bug fixes, no new features - **Pre-releases**: Use -alpha, -beta, -rc suffixes (e.g., 1.0.0-beta.1) - **Build metadata**: +build.123 for CI build info (not part of precedence) - **Commit tags**:
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/ruby-32-with-rbs-steep
______________________________________________________________________ ## priority: high # Ruby 3.2+ with RBS & Steep **Ruby 3.2+ · RBS type definitions · Steep · rbenv · RSpec · Rubocop** - Ruby 3.2+ with .ruby-version file; rbenv for version management - RBS files in sig/ directory parallel to source: lib/foo.rb → sig/foo.rbs - Steep for type checking; avoid Any types, use union and optional types explicitly - RSpec for testing: describe/context/it blocks, 80%+ coverage, function-like test
testing628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/rust-async-await-module-patterns
______________________________________________________________________ ## priority: medium # Rust Async/Await & Module Patterns ## Tokio Runtime Patterns **Single runtime instance** should be created once and shared across the application. ### Pattern 1: Global Runtime (for library bindings) Use `tokio::runtime::Runtime` to manage async code in synchronous contexts (required for FFI bindings). ```rust // src/async_runtime.rs use tokio::runtime::Runtime; use std::sync::OnceLock; static RU
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/testing-philosophy-coverage
______________________________________________________________________ ## priority: critical # Testing Philosophy & Coverage **Three-tier**: unit (80-95%), integration (real DB/services), E2E (smoke/full) **Real infrastructure in tests**: PostgreSQL, Redis, not mocks **Rust**: cargo test, #[tokio::test], 95% coverage (tarpaulin), test error paths, edge cases, no panics **Python**: Function-based tests only (\*_test.py), pytest fixtures, 95% coverage. Structure: tests/{core,features,integra
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/rust-latest-edition-standards
______________________________________________________________________ ## priority: critical # Rust Latest Edition Standards **Rust 2024 edition · High strictness · clippy -D warnings · 95% coverage · Zero unwrap** - Rust edition 2024; cargo fmt, clippy with -D warnings (zero tolerance) - Result\<T, E> for errors; thiserror for custom errors; NEVER .unwrap() in production - Testing: 95% minimum coverage (tarpaulin), unit/integration/doc tests - Documentation: rustdoc on ALL public items with
tools628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/rust-module-organization-public-api-design
______________________________________________________________________ ## priority: high # Rust Module Organization & Public API Design ## Module Structure Principles **Modules are organizational units, not visibility boundaries**. Use `pub` / `pub(crate)` / `pub(super)` to control visibility explicitly. 1. **Root crate module**: `src/lib.rs` or `src/main.rs` re-exports public items 1. **Feature-specific modules**: `src/parsing/`, `src/conversion/`, `src/utils/` organize by domain 1. **Inte
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/rust-polyglot-conventions
______________________________________________________________________ ## priority: critical # Rust Polyglot Conventions **Edition 2024**: let-chains, gen blocks, if/match guards. **Naming**: PascalCase (types), snake_case (fns/vars/modules), SCREAMING_SNAKE_CASE (consts). **Error handling**: Result\<T, Error>, never .unwrap() in production, use `?`, IO errors bubble up properly, SAFETY comments for unsafe code, handle lock poisoning. **Async**: Tokio throughout, #[tokio::main]/#[tokio::tes
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/task-automation-build
______________________________________________________________________ ## priority: critical # Task Automation & Build - Taskfile.yaml for all workflows: setup, dev, lint, format, test, build - Lock files committed: uv.lock, pnpm-lock.yaml, go.sum, Cargo.lock, composer.lock - Dependency minimization: justify all external deps, audit regularly - Version sync across runtimes: Cargo.toml, package.json, pyproject.toml
tools628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/test-apps-published-package-validation
______________________________________________________________________ ## priority: high # Test Apps - Published Package Validation **Purpose**: tests/test_apps validates PUBLISHED/RELEASED packages from npm, PyPI, RubyGems, Maven Central, Docker Hub, and Homebrew. NOT for local development testing. **Location**: tests/test_apps/{python,node,wasm,ruby,go,java,csharp,rust,docker,homebrew,browser-vite-svelte} **Version sync**: Included in `task sync-versions` - automatically updates all test
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/typescript-strictest-standards
______________________________________________________________________ ## priority: critical # TypeScript Strictest Standards __TypeScript 5.x · Strictest typing · No any/object · Generics required · Tests next to source__ - Enable ALL strict flags: strict, noUncheckedIndexedAccess, exactOptionalPropertyTypes - Ban any and object types; use unknown with guards, Record\<string, unknown> - Generics with constraints: <T extends BaseType>, satisfies operator, const assertions - Tests: .spec.ts n
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/elixir-114-with-exdoc-credo
______________________________________________________________________ ## priority: critical # Elixir 1.14+ with ExDoc & Credo **Elixir 1.14+ · Functional-first · ExDoc · ExUnit · Rustler NIF** - Target Elixir 1.14+ with .tool-versions file; use ASDF for version management - Module structure: CamelCase for modules (e.g., Kreuzberg.Native), snake_case for functions/variables - ExDoc for documentation: @moduledoc and @doc tags on all public APIs with examples - ExUnit for testing: describe/tes
tools628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/workspace-dependency-management
______________________________________________________________________ ## priority: critical # Workspace Dependency Management ## Cargo Workspace Fundamentals A workspace coordinates multiple crates under unified configuration. This is **critical** for polyglot projects with core library + language bindings. **workspace/Cargo.toml**: ```toml [workspace] members = [ "crates/html-to-markdown", # Core library "crates/html-to-markdown-py", # PyO3 bindings "crates/h
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/containerization-docker-standards
______________________________________________________________________ ## priority: medium # Containerization & Docker Standards ## Multi-Stage Docker Builds for Rust - **Builder stage**: Compile with full toolchain + dependencies - **Runtime stage**: Minimal base with only compiled binary - **Cache optimization**: Leverage layer caching by copying dependencies before source code - **Size reduction**: 1GB+ builder → 50-100MB runtime image Example Dockerfile for Rust projects: ```dockerfile
tools628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/security-and-vulnerability-management
______________________________________________________________________ ## priority: critical # Security & Vulnerability Management **CVE disclosure · cargo-audit · cargo-deny · Fuzzing · Unsafe code review · Security testing** ## Vulnerability Management - **cargo-audit**: Run on every commit and in CI; fail build on known vulnerabilities ```bash cargo audit cargo audit --deny warnings # Fail on advisories ``` - **cargo-deny**: Detect vulnerable, unmaintained, and unlicensed depe
development628
blockquote-conversion
Blockquote Conversion
tools628
code-block-conversion
Code Block Conversion
development628
emphasis-and-formatting-conversion
Emphasis And Formatting Conversion
development628
form-element-handling
Form Element Handling
tools628
heading-conversion
Heading Conversion
tools628
html-entity-decoding
Html Entity Decoding
development628
image-conversion
Image Conversion
content-media628
inline-code-conversion
Inline Code Conversion
development628
link-conversion
Link Conversion
tools628
list-conversion-and-nesting
List Conversion And Nesting
tools628
markdown-text-escaping
Markdown Text Escaping
development628
markdown-validation
Markdown Validation
tools628
paragraph-conversion
Paragraph Conversion
data-ai628
semantic-html5-element-conversion
Semantic Html5 Element Conversion
development628
table-conversion
Table Conversion
tools628
task-list-detection-and-conversion
Task List Detection And Conversion
tools628
url-sanitization
Url Sanitization
tools628
whitespace-normalization
Whitespace Normalization
tools628
character-encoding-detection
Character Encoding Detection
tools628
input-binary-detection
Input Binary Detection
tools628
text-extraction-and-preservation
Text Extraction And Preservation
tools628
event-handler-removal
Event Handler Removal
tools628
security-configuration-design
Security Configuration Design
testing628
size-limit-enforcement
Size Limit Enforcement
tools628
url-scheme-validation
Url Scheme Validation
tools628
.ai-rulez/skills/conversion-mapping-rules
Conversion Mapping Rules: HTML Elements to Markdown
development628
.ai-rulez/skills/edge-case-handling
Edge Case Handling in html-to-markdown
development628
.ai-rulez/skills/whitespace-handling
Whitespace Handling in html-to-markdown
development628
conversion-mapping-rules
Conversion Mapping Rules: HTML Elements to Markdown
development628
html-parsing-strategies
HTML Parsing Strategies for html-to-markdown
development628
metadata-extraction
Metadata Extraction for html-to-markdown
development628
release-and-deployment-processes
Coordinated multi-platform release and deployment processes with semantic versioning
devops628
visitor-pattern-usage
Visitor Pattern Usage for html-to-markdown
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/agent-selection-usage-guidelines
______________________________________________________________________ ## priority: critical # Agent Selection & Usage Guidelines **When to use agents**: Use the Task tool to spawn specialized agents for focused, language-specific work. Agents provide domain expertise and follow language-specific conventions. **Agent selection rules**: - **Rust core work** → rust-core-engineer (core library, extraction logic, plugins) - **Python bindings** → python-bindings-engineer (PyO3 FFI, Python wrappe
tools628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/binding-crate-architecture-patterns
______________________________________________________________________ ## priority: high # Binding Crate Architecture Patterns ## Overview Binding crates expose Rust libraries to host languages. Each binding framework has distinct patterns, but shared principles apply across all: 1. **Minimal wrapper layer**: Call Rust → host language is glue only 1. **Type translation**: Convert host language ↔ Rust with clear mapping 1. **Error conversion**: Rust errors → native exceptions/error types 1.
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/ci-cd-multi-platform-matrix
______________________________________________________________________ ## priority: critical # CI/CD Multi-Platform Matrix ## GitHub Actions Matrix Strategy Use matrix strategy to test across multiple OS, architectures, and language versions with minimal duplication: ```yaml name: CI on: push: branches: [main, develop] pull_request: jobs: test-matrix: runs-on: ${{ matrix.os }} strategy: fail-fast: false matrix: include: # Linux: x86_64 and A
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/cicd-pipeline-standards
______________________________________________________________________ ## priority: critical # CI/CD Pipeline Standards **Architecture**: Stages: Validate (lint/format) → Build → Test (unit/integration) → Deploy. Quality gates: zero warnings, tests pass, coverage thresholds met. **Multi-platform Testing**: linux/amd64, linux/arm64, macOS (Intel/ARM). Docker: multi-stage builds, minimal base images (alpine/distroless). **Artifact Management**: Cache dependencies (Cargo, npm, Maven, Go module
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/common-task-commands
______________________________________________________________________ ## priority: critical # Common Task Commands **Setup & Installation**: - `task setup`: Install all dependencies (Rust, Python, Node, Go, Java, Ruby, etc.) - `task setup-pre-commit`: Configure pre-commit hooks **Build Commands**: - `task build`: Build all (respects BUILD_PROFILE) - `task build:all`: Build all languages - `task build:all:dev`: Build all in debug mode - `task build:all:release`: Build all in release mode -
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/core-principles
______________________________________________________________________ ## priority: critical # Core Principles **Do only what's asked. Never create files unnecessarily. Prefer editing. No proactive docs/READMEs.** **Python**: Builtin imports at top, dataclasses frozen/hashable/slots, function-based tests only. **Rust**: Never .unwrap() in production, SAFETY comments for unsafe, handle lock poisoning. **Architecture**: ALL extraction logic lives in Rust core. Bindings provide language-idiom
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/documentation-standards-language-parity
______________________________________________________________________ ## priority: critical # Documentation Standards & Language Parity **CRITICAL: Full language parity required**. ALL documentation, guides, and code examples MUST include snippets for ALL 7 supported languages: Rust, Python, TypeScript, Ruby, Java, Go, C#. **Code snippet structure**: - Location: docs/snippets/{language}/{category}/{filename}.{ext} - Categories: getting-started, config, advanced, ocr, metadata, plugins, api
tools628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/elixir-bindings-patterns
______________________________________________________________________ ## priority: high # Elixir Bindings Patterns (Rustler NIF) **Role**: Elixir bindings for Rust core using Rustler NIF bridge pattern. **Scope**: - Rust NIF bridge: packages/elixir/native/<project>\_rustler/ (Rust crate with cdylib output) - Elixir wrapper: packages/elixir/lib/<project>/ (OTP application with public API) - ExUnit tests: packages/elixir/test/ **Architecture**: Elixir OTP application → Rustler NIF (.so/.dyl
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/error-handling-strategy
______________________________________________________________________ ## priority: critical # Error Handling Strategy **CRITICAL: OSError/RuntimeError must ALWAYS bubble up** (Python + Rust). SystemExit, KeyboardInterrupt, MemoryError too. **Python**: Exception-based, inherit from KreuzbergError. OSError patterns: 1) Library misuse→bubble up, 2) Subprocess→analyze stderr for parsing keywords, 3) Cache→ignore, 4) Dependencies→MissingDependencyError or bubble up. Always add ~keep comments. *
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/fixture-driven-testing-strategy
______________________________________________________________________ ## priority: high # Fixture-Driven Testing Strategy ## Shared Test Fixtures Across Language Bindings - **Single source of truth**: Fixtures defined once in Rust, exposed to all language bindings - **API parity validation**: Identical behavior across languages guaranteed by fixtures - **Fixture generation**: Rust generates JSON/YAML fixtures from canonical implementation - **Language-specific consumption**: Each binding co
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/git-commit-standards
______________________________________________________________________ ## priority: critical # Git & Commit Standards - Conventional Commits 1.0.0: feat/fix/docs/refactor/test/chore with scope - NEVER include AI signatures ("Generated with Claude") in commits - Pre-commit hooks with prek/lefthook/husky: linting, formatting, tests - Branch protection: main/development with required status checks
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/go-bindings-patterns
______________________________________________________________________ ## priority: high # Go Bindings Patterns **Role**: Go bindings for Rust core. Work on CGO bridge and Go SDK/E2E suite. **Scope**: Go 1.25 module, CGO wrappers around C FFI, Go-idiomatic config/result structs, linting setup, benchmark harness scripts. **Commands**: cd packages/go/v4 && go test ./..., golangci-lint run ./..., ensure `LD_LIBRARY_PATH`/`DYLD_LIBRARY_PATH` includes target/release when running tests. **FFI**:
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/modular-taskfile-structure
______________________________________________________________________ ## priority: critical # Modular Taskfile Structure **Root**: Taskfile.yml (version 3) includes all modular task files from .task/ directory. **Configuration Files**: - `.task/config/vars.yml`: Global variables (BUILD_PROFILE, VERSION, PDFIUM_VERSION, ORT_VERSION, GOLANGCI_LINT_VERSION, paths, OS/ARCH detection, CARGO_PROFILE_DIR mapping) - `.task/config/platforms.yml`: Platform-specific variables (EXE_EXT, LIB_EXT, NUM_C
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/monitoring-observability-standards
______________________________________________________________________ ## priority: medium # Monitoring & Observability Standards ## Structured Logging with Tracing Crate - **Use `tracing` crate**: Unified logging, tracing, and metrics in Rust - **Structured events**: key=value pairs instead of f-string format - **Span context**: Automatic propagation of request IDs, user info, etc. - **Multiple subscribers**: Layer logs, metrics, and distributed traces together Basic setup in Rust: ```rus
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/platform-support-cross-platform-compatibility
______________________________________________________________________ ## priority: critical # Platform Support & Cross-Platform Compatibility **Supported Platforms**: Windows (via Git Bash/MSYS2/MinGW/PowerShell), Linux (all major distributions), macOS (Intel & ARM/M-series). **Platform Detection in Taskfiles**: - `{{.OS}}`: Detects operating system (windows, linux, darwin) - Windows detection: PowerShell, GOOS env, WINDIR/SYSTEMROOT, MSYSTEM (Git Bash/MSYS2/MinGW) - Linux/macOS: uname
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/polyglot-bindings
______________________________________________________________________ ## priority: medium # Polyglot Binding Architecture ## Language Binding Pattern Each binding crate (Python, TypeScript, Ruby, PHP, Go, Java, C#) follows: 1. **Minimal wrapper layer**: Call Rust functions directly, no business logic 1. **Type translation**: Convert host language types ↔ Rust types 1. **Error mapping**: Rust errors → language-native exceptions 1. **Documentation**: Link bindings to Rust docs, add language-
development628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/project-structure
______________________________________________________________________ ## priority: medium # Project Structure & Conventions ```text html-to-markdown/ ├── crates/ │ ├── html-to-markdown/ # Core Rust library (main logic) │ ├── html-to-markdown-py/ # PyO3 bindings for Python │ ├── html-to-markdown-node/ # NAPI-RS bindings for Node.js │ ├── html-to-markdown-rb/ # Magnus bindings for Ruby │ ├── html-to-markdown-php/ # ext-php-rs extension for PHP │ ├
tools628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/python-bindings-patterns
______________________________________________________________________ ## priority: critical # Python Bindings Patterns **Role**: Python bindings for Rust core. Work on PyO3 bridge and Python wrapper packages. **Scope**: PyO3 FFI, Python-idiomatic API, Python-specific extensions, postprocessors. **Commands**: maturin develop, pytest, ruff format/check. **Critical**: Core logic lives in Rust. Only Python code for bindings, Python-specific extensions, or API wrappers. If core logic needed, c
tools628
.remote-cache/kreuzberg-shared-rules/.ai-rulez/skills/quick-start
______________________________________________________________________ ## priority: medium # Developer Quick Start ## Prerequisites - Rust 1.75+ (stable or nightly for WASM) - Python 3.10+ with uv package manager - Node.js 18+ with pnpm ≥10.17 - Ruby 3.2+ with rbenv - PHP 8.2+ with Composer - Go 1.25+ (optional, for Go binding) - Java JDK 22+ (optional, for Java binding) - .NET 8.0+ SDK (optional, for C# binding) - Task (task runner) - prek (pre-commit hook manager) ## Quick Setup ```bash
development628