Browse reusable Agent Skills, each with a clear purpose and practical guidance.
Creates Rails concerns for shared behavior across models or controllers with TDD. Use when extracting shared code, creating reusable modules, DRYing up models/controllers, or when user mentions concerns, modules, mixins, or shared behavior. WHEN NOT: Logic used by only one model or controller (keep it in place), complex business logic (use service objects), or query encapsulation (use query objects).
Configures Solid Queue for background jobs in Rails 8. Use when setting up background processing, creating background jobs, configuring job queues, or migrating from Sidekiq to Solid Queue. WHEN NOT: Synchronous in-request processing, real-time WebSocket features (use Action Cable), or simple operations that don't need background execution.
Behavioral guidelines to reduce common LLM coding mistakes. Use when writing, reviewing, or refactoring code to avoid overcomplication, make surgical changes, surface assumptions, and define verifiable success criteria.
Feature development with cross-module impact analysis. Covers trainer hierarchy, model adapters, reward pipeline, config system, sample dataclasses, and distributed training paths. Trigger: 'add feature', 'implement', 'refactor', 'reorganize', 'new capability'.
Complete workflow for adding a new RL training algorithm. Covers paradigm selection, TrainingArguments subclass, trainer implementation, registry, example config, and verification. Trigger: 'add algorithm', 'new trainer', 'new training method', 'implement algorithm'.
Complete workflow for adding a new model adapter. Covers analysis, sample dataclass, adapter implementation (4 abstract methods + per-modality encoder overrides), registry, example YAML, and verification. Trigger: 'add model', 'support new model', 'integrate model', 'new adapter'.
Complete workflow for adding a new reward model. Covers pointwise vs groupwise design, __call__ contract, registration, YAML config, multi-reward setup, and verification. Trigger: 'add reward', 'new reward model', 'custom reward', 'scoring function'.
Resolve pull request code review comments end-to-end. Fetches inline review threads from a GitHub PR URL, verifies or implements fixes in the working branch, resolves threads, and posts a summary reply. Use when the user shares a PR link, asks to address/fix/resolve code review feedback, or mentions gemini-code-assist, Copilot, or Bugbot review comments on a pull request.
Enforces strict avoidance of unwrap() in Rust code. Use when writing, editing, or reviewing Rust (.rs) code in this project. Prevents unsafe panic-prone patterns and promotes proper error handling alternatives.
Use this skill whenever the user asks to regenerate the S2S API artifacts, refresh the server swagger output, or update the generated admin-panel auth API client from the backend schema and routes. Also use this skill proactively whenever the current task changes any S2S generation inputs, even if the user did not explicitly ask for regeneration. Relevant changes include updates under `server/src/routes/`, `server/src/scripts/schemas/`, `server/src/scripts/generate-swagger.cjs`, or `admin-panel/scripts/generateService.ts`. Trigger on requests like "generate s2s", "regenerate swagger", "refresh auth api", "update the generated s2s client", or "run s2s:generate", and also when those source files were edited as part of the work.