全部能力

Anthropic's method for training harmless AI through self-improvement. Two-phase approach - supervised learning with self-critique/revision, then RLAIF (RL from…

Builds and launches Docker containers with health verification. Use when validating that containerized services start correctly.

Configures structured JSON logging with Serilog (.NET) or structlog (Python). Use when adding logging to backend projects.

Educational GPT implementation in ~300 lines. Reproduces GPT-2 (124M) on OpenWebText. Clean, hackable code for learning transformers. By Andrej Karpathy.…

Sets up test infrastructure with Vitest, xUnit, and pytest. Use when adding testing frameworks and sample tests to a project.

Generates .NET Clean Architecture backend structure from entity definitions. Use when bootstrapping .NET backend projects.

Audits whether test results can be trusted: flakiness, isolation, real external dependencies, time/random/order dependency, and shared state. Use when auditing…

Checks layer leakage in method signatures, missing DTOs, entity leakage to API, inconsistent error contracts. Use when auditing API contracts.

Story-level quality gate with 4-level verdict (PASS/CONCERNS/FAIL/WAIVED) and Quality Score. Use when Story is ready for quality assessment.

Scaffolds new React projects or restructures monoliths to component-based architecture. Use when setting up frontend structure.

Web research skill using Tavily API for search, extract, crawl, map, and structured research tasks. Use when you need latest information, extract content from…

Verifies bootstrapped projects via build, test, and container health checks. Use when validating project setup completeness.

Executes Story tasks in priority order (To Review, To Rework, Todo). Use when Story has planned tasks ready for implementation.

Sets up logging, error handling, CORS, health checks, and API docs. Use when adding cross-cutting concerns to backend projects.

Comprehensive healthcare AI toolkit for developing, testing, and deploying machine learning models with clinical data. This skill should be used when working…

Fixes tasks in To Rework by applying reviewer feedback, then returns to To Review. Use when task was rejected during review.

Sets up linters, pre-commit hooks, and test infrastructure. Use when adding code quality tooling to a project.

Document deployment processes, infrastructure setup, CI/CD pipelines, and configuration management. Use when creating deployment guides, infrastructure docs, or CI/CD documentation.

Checks compiler/linter errors, deprecation warnings, type errors, failed tests, build config issues. Use when auditing build health.

Scores each test by portfolio value and returns KEEP/DELETE/MERGE/REWRITE. Use when pruning test-suite cost.