技能列表

>

Turn rough product, feature, or repo-change requests into a PM-grade plan plus agent-ready execution context. Use whenever the user has a vague idea, wants a PRD/spec/phase plan, needs repo-aware context for Codex/Antigravity/OpenClaw, wants a large request cut into a clear MVP or phased plan, or asks for a coding-agent handoff that must preserve scope, constraints, and acceptance criteria. Prefer this skill over generic brainstorming or direct coding when the work needs product framing before o

Persistent cross-session memory, structured observability, encrypted secrets management, and replay for OpenClaw agents. Local-first SQLite. Installs as both skill and OpenClaw plugin. Use when: (1) agents need memory between sessions, (2) API keys need secure storage, (3) run history needs replay and analysis, (4) auth failures need classification and recovery.

>

解决AI记忆混乱问题——"今天讲明天忘"。通过文件系统和规则流程实现持久记忆，确保新任务开始时自动回顾上下文，而不是靠"想"。核心机制：任务开始判断→读取记忆→执行→保存。适用于频繁切换话题或任务的AI协作场景。

AI-powered context management for OpenClaw sessions

Proactive context monitoring with smart 3-level alerts. Know when to restart before quality degrades.

>

Use this skill whenever the conversation context is getting long, when a user asks to "compress", "summarize", or "clean up" the conversation, or when you detect the context window is filling up. Also triggers automatically via PreCompact hook if configured. Compresses conversation history using a tiered strategy — preserving what matters, summarizing what's useful, dropping what's noise — then writes a structured memory file so nothing important is truly lost. Use this even for partial compress

>

Use this skill when managing long-context conversations and tasks. Provides intelligent compression, key extraction, context enhancement, memory retrieval, and context prediction to optimize context window usage and maintain conversation coherence.

|