Kunlun Yaochi AI Precision Memory Management
Paste the following prompt into your AI chat to install this skill:
Please install @syln/kunlunyaochi following https://skillhub.cn/install/skillhub.md.
About this skill
What problem it addresses
When an AI agent runs long-term, important conclusions, rules, and task state often scatter across conversations, temporary files, and local docs. After reinstall, migration, or context loss, the agent may not recover a coherent memory state from the current session alone. KLYC-PMM targets this case: it writes text knowledge into recoverable, searchable, synchronizable memory entries and keeps long-term hooks in MEMORY.md. It is not a general vector database or ETL tool; its focus is persistent agent memory and disaster recovery.
How the skill works
After initialization, the agent obtains a Kunlun URL, which acts as the main credential for restoring memory. Daily writes use push and require an explicit memory domain such as public, shared, essence, private, or disaster; data is compressed with Brotli, encrypted through multiple layers, and uploaded to the Yaochi service. Recovery uses recover, local lookup uses search, and remote lookup uses search-yaochi.
Watch mode monitors files such as MEMORY.md, batches file changes, and pushes updates within a window. hooks-pull periodically fetches distilled hooks and writes them into a bounded section of MEMORY.md. self-test provides idempotent environment checks, while hook-check compares local and remote hook states. Paid tiers such as huhunfu and fenshenfu require a payment path; if the agent lacks a payment plugin, the user can complete payment instead.
Boundaries and caveats
The solution depends on curl, jq, and HTTPS connectivity, so it is unsuited to fully offline environments or millisecond-level read/write workloads. It is intended for text knowledge rather than binary media such as images, audio, or video. If the Kunlun URL is lost, it cannot be recovered, so it must be stored securely; MEMORY.md may also contain recovery credentials and should not be committed to public repositories. For temporary notes, one-off batch jobs, or large binary storage, using this skill adds unnecessary encryption, sync, and distillation overhead.
Use Cases
- After reinstall, restore core rules from the Kunlun URL.
- Push key conclusions to private domain, then search locally.
- Watch MEMORY.md changes and sync updates to remote memory.
- Run hook-check to flag stale distillation hooks.
Best For
- Agent engineers who need AI memory recovery after migration.
- Developers needing local agent search plus cloud backup.
- Project leads auditing multi-agent distillation hook health.
- SREs scripting agent file watching and disaster recovery.
Related Skills
An AI workflow that structures personal knowledge bases, IMA, Yuque, and Feishu content into approved personas for paid WeChat mini-agent deployment.
Local workflow memory with matching and SOP updates.
An OpenClaw live streaming executor that initializes TRTC streaming, starts a real-time dashboard, generates viewer URLs, and continuously reports live events.
Breaks down physical supply chains for super-trends to identify second- and third-layer bottlenecks, runs valuation and reverse checks, and maintains trackable reports.