adr-management
Create, index, and maintain Architecture Decision Records (ADRs) using the Michael Nygard template.
#reversible
39 skills tagged reversible · risk axis.
api-documentation
Generate, validate, and maintain API endpoint documentation including OpenAPI specs, request/response examples, and versioned API references.
api-implementation
Implement API handlers/controllers based on an existing API contract (OpenAPI spec). Generates route handlers, validation, and error handling.
caching-strategy
Design and implement caching layers using Redis, Memcached, or in- memory caches with proper invalidation strategies.
changelog-generation
Generate changelogs from commit history using semantic versioning and conventional commits, including release notes and breaking change notices.
ci-pipeline
Generate CI/CD pipeline configurations for GitHub Actions, GitLab CI, or Azure Pipelines.
cleanup
Remove noise, enforce formatting, and safely rename identifiers. A mixed-mode skill: cosmetic fixes run in Eco, renaming runs in Power.
code-generation
Generate production-ready code from specifications, templates, and domain models. Enforces language idioms, DRY principles, and project conventions.
data-access
Implement data access layers using the Repository pattern. Prevents N+1 queries, manages transactions, and optimizes reads.
db-tuning
Database performance tuning: index analysis, query plan optimization, connection pooling, and configuration tuning.
decision-records
Create, manage, and query Architectural Decision Records (ADRs) using the context/decision/consequences format to capture design rationale.
dependency-updates
Automate and manage dependency updates with semver-aware compatibility checks, breaking change detection, and lockfile management.
docker-containerization
Write optimized Dockerfiles with multi-stage builds, security best practices, and minimal image sizes.
error-handling
Standardize error types, logging formats, and error response structures across the project.
error-handling-debug
Debug and improve error handling strategies including exception hierarchies, retry logic, graceful degradation, and circuit breakers.
feature-shipping
Master workflow for shipping a feature end-to-end. Orchestrates PRD authoring, task decomposition, TDD implementation, code review, and the release pipeline as a single named sequence ending in an audit. Contains no implementation logic.
git-workflow
Guide branching strategies, commit conventions, merge vs rebase decisions, and conflict resolution workflows for consistent version control practices.
inline-documentation
Analyze, generate, and improve inline code documentation including JSDoc, docstrings, type annotations, and comment quality.
integration-testing
Test across component boundaries using real databases, HTTP clients, and message queues via Docker containers.
knowledge-sharing
Establish documentation culture, wikis, runbooks, onboarding guides, and tribal knowledge capture to reduce bus factor and accelerate teams.
monitoring-setup
Design and configure observability stacks covering metrics, logs, traces, alerting, SLIs/SLOs, and dashboards.
mutation-testing
Verify test suite quality by injecting faults (mutants) into the code and checking if tests catch them.
openapi-specs
Generate or update OpenAPI (Swagger) specifications from existing code, or validate code against existing specs.
pipeline-context
Pre-compute expensive operations once and pass via shared context object to all pipeline stages. Eliminates O(n) recomputation in multi-stage processing.
pr-management
Manage pull request lifecycle including templates, reviewer assignment, merge criteria, and stale PR handling for efficient code integration.
prd
Adaptive PRD authoring — selects the right format (lean 1-pager, full structured, working backwards, or hypothesis-driven), guides section-by-section completion, and manages document lifecycle with versioning, status transitions, and amendment logging.
progressive-execution
Return fast approximate results immediately (Phase 1), then enhance with slower enriched results (Phase 2). Phase 1 is standalone. Phase 2 is optional enhancement. If Phase 2 fails, Phase 1 stands.
readme-generation
Create comprehensive README files with installation instructions, usage examples, API reference, and contribution guidelines.
refactoring
Safely restructure code while preserving external behavior. Applies proven refactoring patterns with mandatory test verification.
refactoring-suite
Complex refactoring operations with strict test verification. Every refactor follows: Test → Refactor → Test → Revert-on-Fail.
refactoring-workflow
Master workflow for refactoring at four named depth levels — cosmetic, micro, meso, architectural — that compose as a ladder. Always runs chesterton-fence and style-conformance up front; always closes with a characterisation/test safety net, minimal-diff scope checks, and the auditor. Contains no implementation logic; orchestrates only existing category skills.
release-notes
Parse git history to generate changelogs and release notes following Conventional Commits and Keep a Changelog formats.
repo-maintenance
Adaptive cleanup framework for repo-wide structural optimization. Embodies the principle: start simple, explore, discover value, pivot strategy toward the greater good. Deletion is the last resort. Cleanup is the example — adaptive direction-seeking is the lesson.
resilience-patterns
Graceful degradation when external dependencies fail. Circuit breaker state machine, retry with backoff, bulkhead isolation, fallback chains, and timeout management.
responsive-design
Make a web page work on every screen — phone to ultrawide — without text or media ever escaping its container. Invoke whenever building or changing any page/CSS that will be viewed on more than one screen size (i.e. always). Overflow and word-wrap bugs are treated as correctness bugs, not polish.
schema-driven-config
Define configuration as a single schema that powers defaults, validation, storage, and optional UI generation. One entry per setting, zero duplication.
scorer-pipeline
Decompose complex evaluation into independent, composable micro-scorers with shared context and explicit weights. Universal pattern for any multi-factor ranking or assessment system.
tdd-workflow
Drive implementation through the Red-Green-Refactor cycle. Write failing tests first, implement minimal code to pass, then refactor.
unit-testing
Generate unit tests following AAA pattern (Arrange-Act-Assert). Supports Jest, JUnit, Go Test, PyTest, and more.