按命令族和 profile 选择合适的 OpenClaw CLI 命令,并校验执行结果。
编程
OpenClaw Backstory Builder
试用Build, bootstrap, and maintain a stable OpenClaw default agent identity by interviewing the user, updating IDENTITY.md, and seeding core context files (AGENT...
它能做什么
Build, bootstrap, and maintain a stable OpenClaw default agent identity by interviewing the user, updating IDENTITY.md, and seeding core context files (AGENT...
技能文档
OpenClaw Default Agent Backstory
Use this skill to define or refine a default OpenClaw agent identity and its supporting context files.
Reference model: OpenClaw context and bootstrap behavior from https://docs.openclaw.ai/concepts/context.
Core Rules
- Ask five questions every time unless the user explicitly asks to skip, shorten, or focus.
- Use ranked questions from
references/question-bank.md. - Prefer unanswered or weakly defined areas first.
- Keep content practical and stable for daily use.
- Preserve user-authored content unless the user asks for replacement.
- If user asks for a default memory policy, apply the
Primary Daily Driverpolicy inreferences/bootstrap-core-files.md.
Mode Selection
Select one mode before writing files:
bootstrapmode:
- Use when user asks to bootstrap/onboard/start fresh.
- Use when core files are missing or mostly placeholders.
- Create or refresh baseline content for all core context files.
identity-refreshmode:
- Use when user asks for personality/backstory updates only.
- Update
IDENTITY.mdsections without rewriting unrelated core files.
Core Context Files
In bootstrap mode, ensure these files exist at workspace root with usable starter content:
AGENTS.mdSOUL.mdTOOLS.mdIDENTITY.mdUSER.mdHEARTBEAT.mdBOOTSTRAP.md
Use guidance and starter structure in references/bootstrap-core-files.md.
Also ensure memory scaffolding exists:
MEMORY.mdmemory/directory for daily notesMEMORY.mdmust be a standalone workspace-root file, not a section insideAGENTS.md.
Bootstrap Interview Flow
- Read existing core files if they exist.
- Ask exactly five bootstrap-priority questions from
references/question-bank.md. - Synthesize answers into practical defaults.
- Write or refresh all core files listed above.
- Set rational session-management and memory defaults:
- session startup read order
- when to create/update daily memory notes
- what belongs in
MEMORY.mdvs daily files - heartbeat cadence and quiet-hour behavior
- Keep
BOOTSTRAP.mdshort, with current setup status and next steps. - Summarize what was created, what was updated, and any remaining gaps.
Identity Refresh Flow
- Read
IDENTITY.md. - Ask exactly five identity-priority questions from
references/question-bank.md. - Update
IDENTITY.mdwith:
## Backstory## Behavioral Guardrails## Growth Arc
- Map each update directly to user answers.
- Summarize changes and open questions.
Writing Constraints
- For
IDENTITY.md, preserve existing header fields unless asked to change them. - Keep sections concise and concrete.
- Explicitly document boundaries/refusals.
- Do not invent personal history if user input is missing; mark assumptions clearly.
Completion Criteria
Finish only when:
- five questions were asked (unless user requested otherwise), and
- required target files for the selected mode were updated with concrete, non-placeholder content.
相关技能
Configure a powerful macOS machine to host multiple dedicated OpenClaw agent users with Fast User Switching, a shared Homebrew toolchain, per-user OpenClaw h...
Allows this local OpenClaw agent to engage and socialize on the global OpenClaw Community Social Network.
Use this skill when the user says "你好世界" / "你好世界初始化" / "初始化龙虾" / "龙虾初始化" / "初始化" / "bootstrap" / "hello world" / "onboarding" / "重新认识一下" / "重置我的偏好" / "更新画像",...
Audit an OpenClaw agent workspace and generate standardized evaluation reports, scores, and patches. Use when asked to review memory quality, retrieval effic...
Create and wire a new OpenClaw agent with a fixed workflow. Use when the user asks to create/add a new OpenClaw agent or says “我要创建一个新的 Agent”, automate mult...