Design, explore, implement, polish, redesign, and review Epicenter interfaces from product direction through component-system collapse. Use when co-designing a screen, asking "what do other apps do", comparing UI directions or comparable apps, turning a rough feature idea into a buildable Svelte surface, choosing @epicenter/ui components, changing packages/ui, or replacing class-heavy markup and local interface primitives. Not for a tiny CSS-only repair unless it reveals a broader pattern.
Present proposed code changes visually before implementing: before/after diffs, ASCII diagrams, comparison tables. Use when: "show me options for this change", "compare approaches", or a multi-file change needs a before/after comparison before editing. Not for planning questions with no concrete code change on the table.
Co-create an explicitly accepted, uncompromising destination before planning or implementation. Use when the user wants to discover what they actually want, asks for an uncompromising vision, wants iterative synthesis and pushback, is exploring an architecture or product model before a plan exists, or says to use dialectic. Do not use for interrogating an existing plan, comparing one bounded implementation choice, or ordinary implementation with a settled destination.
In-code documentation, folder READMEs, code comments. Use when: "document this", "add JSDoc", "write a README", "explain this code", or writing README.md/JSDoc.
Draft a compact, cold-start prompt to paste into Claude Code, with Claude as orchestrator and Codex available through /codex:rescue for bounded evidence gathering, focused implementation, commands, verification, cleanup, and creative support. Use when the user says "hand this off", "wrap up for the next session", "resume this later", "continue in a new chat", "write a continuation prompt", "make a prompt I can copy-paste", "create a delegation brief", "prompt for an orchestrator", or invokes /handoff. Not for /goal lines (use agent-goal) or prompts that ship inside product code.
Write technical specs that let agents implement autonomously. Use for "write a spec", "plan this feature", "create a planning doc".
Technical articles and blog posts with honest trade-offs. Use when: "write a blog post", "draft an article", "write about this", creating articles in docs/articles/.
Create, revise, place, audit, evaluate, and validate repository agent instructions across AGENTS.md, CLAUDE.md, and `.agents/skills`. Use when adding or moving agent rules, deciding whether guidance belongs globally or in a skill or reference, writing or improving a skill, tuning routing descriptions, validating discovery, or reviewing instruction sprawl and composition.