메뉴
Create a slash-command `/goal` for long-running Codex or Claude Code work when the user explicitly asks for a `/goal`, agent goal, or completion condition. Outputs one goal line with the objective, starting context, validation evidence, and stop condition.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Run a continuous collapse-and-simplify pass that surgically removes indirection failing to earn its boundary. Use when the user says 'collapse pass', 'simplify pass', 'reduce indirection', 'shrink the surface', 'find what to delete', when asking to audit a package for dead abstractions, when reviewing a pull request, branch, or recent merged change for simplification (isolated in a worktree), or when the goal is a sequence of small refactor commits that delete more than they add. Pairs with code-audit (smell catalog), refactoring (per-change mechanics), one-sentence-test (cohesion gate), cohesive-clean-breaks (deep redesigns), approachability-audit (first-read sanity), and post-implementation-review (second-read after each commit).
How a workspace-backed app under `apps/*` is composed: the isomorphic doc factory (`create<App>`), the environment factories (`open<App>Browser` / `open<App>Extension` / tauri), the `#platform/*` build-time platform DI for multi-platform (Tauri) apps, the `session` singleton, daemon/script placement under per-project `workspaces/<app>/`, and the file layout itself. Use when creating a new app, naming or placing the iso/browser/extension factory, wiring `#platform/*` subpath imports for a Tauri seam, choosing between auth-gated (Shape A) vs module-singleton (Shape B), placing the session singleton, or registering daemon/script bindings.
Query/RPC layer with TanStack Query, defineKeys, service composition, runtime DI. Use for createQuery, createMutation, queries/mutations, reactive data management.
TanStack AI patterns for @tanstack/ai, @tanstack/ai-svelte, chat state, streamed responses, UIMessage parts, tool calling, tool approvals, and provider model adapters. Use when working on AI chat, createChat, fetchServerSentEvents, UIMessage conversion, or TanStack AI tools.
List 3-5 comparable apps when planning a UX surface to test category fit and surface high-leverage refusals. Use for "what do other apps do", identity, sync state, local-first design.
Draft a self-contained, copy-pasteable prompt so a fresh agent can continue without this thread. Use when the user says "hand this off", "compact this", "wrap up for the next session", "write a continuation prompt", "draft a prompt", "make a prompt I can copy-paste", "create a delegation brief", or invokes /handoff.
| name | agent-goal |
| description | Create a slash-command `/goal` for long-running Codex or Claude Code work when the user explicitly asks for a `/goal`, agent goal, or completion condition. Outputs one goal line with the objective, starting context, validation evidence, and stop condition. |
Write one /goal line that a coding agent can keep pursuing until the transcript proves the work is done.
This skill owns the slash-goal artifact. Do not use it for ordinary prompt drafting, handoffs, continuation prompts, delegation briefs, or prompt engineering.
/goal <single objective> in <lane> until <observable evidence proves completion>. First read <starting context>. Work in checkpoints and surface <validation evidence> after each checkpoint.
Answer these in order, then compress them into one line:
What should change?
Where should the agent start?
What evidence proves it changed?
When should the agent stop?
Include only execution-critical detail:
/goal when the user asks for the goal text.bun test packages/auth exits 0" beats "tests pass."/goal for vague wishes, open-ended research, unrelated chores, or work with no observable evidence.Different agents may evaluate goals differently, but the portable rule is the same: the goal should not rely on hidden state. Tell the agent to run checks and surface evidence in the transcript.
Before handing back a goal, verify: