oli-torus

Clean up and harden code introduced by the current branch without drifting into broad refactors. Use when Codex should review branch-local changes against a base branch, improve newly added or modified code for reuse, clarity, docs, specs, and dead-state cleanup, and optionally work interactively with the user in `self-driving`, `smart`, or `assisted` mode. This skill is for targeted cleanup after feature iteration, especially when Elixir, Phoenix LiveView, or React changes may have left duplicated helpers, missing docs/specs, stale assigns/props/attrs, or overlapping state.

Convert a provided UI design source into an implementation-ready design brief that maps visual intent to Torus design tokens, icons, reusable components, and code targets. Use when a Jira ticket or developer provides Figma or another design reference and the team needs governed implementation guidance before coding. Supports durable feature briefs and lightweight ticket-level briefs.

Execute the repo-local Figma-backed UI implementation workflow, including governed design mapping, current-state audit, targeted implementation, verification, iteration, and runtime memory management outside the repository.

Triage and address pull request review comments after a PR is open. Use when Codex should read the linked ticket if present, the PR description, the relevant Torus review guidelines, and all review comments to build context first; then classify comments internally, review them interactively with the user one by one, agree on an action per comment, implement only the approved changes, create a single follow-up commit, and reply thread-by-thread for anything deferred, clarified, or not taken.

Author and maintain Oli.Scenarios YAML-driven integration tests for Torus, including new `.scenario.yaml` files and companion ExUnit runners. Use when features or bugs need non-UI integration coverage through real authoring/delivery/student workflows (projects, content edits, publishing, sections, enrollments, learner attempts, assertions), especially when other `spec_*` skills need scenario-based test implementation. Do not use for UI/browser automation or for tests that depend on fixtures/factories/mocks for domain setup.

Extends Oli.Scenarios infrastructure when required feature coverage is not yet supported by current directives. Use for adding new directive capabilities, parser/validator/schema support, handler execution paths, and infrastructure tests so downstream scenario tests can be authored via `build_scenario`.