zenstory-ai/oh-story-claudecode / skills/story-setup
story-setup
Flagged网文写作工具集基础设施部署与检查。为 Claude Code / OpenCode / Codex / Google Antigravity / ZCode / OpenClaw / Reasonix 提供内置适配;Web AI / 通用 Agent 可走 skills + AGENTS.md 文件模式。触发方式:/story-setup、$story-setup、「准备写书」「帮我搭一下环境」「配置写作项目」「检查写作环境」。
references/antigravity/rules/oh-story.md
View on GitHub---
trigger: always_on
---
# oh-story writing project rules
This workspace uses the oh-story web-fiction skills in `.agents/skills/`.
Read the selected skill's `SKILL.md` first, then load references as it directs.
## Routing
- Long-form writing or continuation: `story-long-write`
- Short-form writing: `story-short-write`
- Long/short deconstruction: `story-long-analyze` / `story-short-analyze`
- Long/short market scan: `story-long-scan` / `story-short-scan`
- Remove AI-writing patterns: `story-deslop`
- Adversarial review: `story-review`
- Import an existing story: `story-import`
- Cover generation: `story-cover`
- Publishing preparation (book title, synopsis, platform tags): `story-publish`
- Ambiguous story intent: `story`
- Project deployment/update: `story-setup`
- Reuse an authenticated Chrome session: `browser-cdp`
## Writing guardrails
- Keep every story artifact, temporary drafting segment, tracking transaction,
and generated project directory inside the current workspace. Never create or
continue an oh-story book under `~/.gemini/`, Antigravity's `scratch/`, or any
other directory outside the workspace unless the user explicitly names that
external destination.
- Before writing prose, long-form projects require the matching
`大纲/细纲_第N章*.md`; short-form projects require `小节大纲.md`.
- Treat `追踪/_tracking-state.json` as the structured source of truth. Do not
hand-edit its derived Markdown views.
- After prose is written, resolve every deterministic finding injected by the
oh-story Antigravity hooks before continuing to another chapter.
- Prefer the seven deployed custom agents in `.agents/agents/` for specialist
work. Call `invoke_subagent` with `TypeName`, a concise `Role`, and `Prompt`
(include concrete file paths in `Prompt` so read-only subagents in `agy` CLI
can read them directly via `read_file`/`view_file` when `find_by_name` or
`grep_search` is unavailable). Because `invoke_subagent` runs asynchronously
in the background, end the turn to wait for the subagent's `send_message`
wakeup before running dependent steps such as `manage_analysis_run.js commit`
or review synthesis; if the runtime cannot start custom subagents, use the
skill's documented solo/direct fallback instead of failing the workflow.
## Working with the author
- Use plain story terms (book, chapter, outline, setting) in the main reply.
Quote script, field, or status names when reporting an error.
- Skill templates marked `<!-- author-report -->` are format references only:
reply with the text inside, never the code fence.
- When the author asks to remember, forget, or confirm a writing habit or
preference, use author memory management in `story-long-write` or
`story-short-write`, based on the author's explicit type or the active book.
For a general habit with no active book, use the long-form entry. Ask which
book only when a book-specific preference has no identifiable target; do not create a book.
Write `.story/作者记忆/` in the relevant workspace or book directory. Never store it in Antigravity's
built-in memory. One-off requests are simply carried out, not recorded.
- Never modify installed skill files (`SKILL.md`, `references/`, `scripts/`) in
this workspace (redeploying via `story-setup` is the exception). If a skill script fails, stop and report its error and command; do not patch around it.
## Context recovery
At the start of a new conversation, and whenever context appears compacted or
story state is uncertain, locate the active book and read `追踪/上下文.md`
(long-form) or `设定.md` + `小节大纲.md` (short-form) before continuing. The
Antigravity external hook API has no PreCompact/PostCompact event, so this rule
is the mandatory recovery path after compaction.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |