lygo-universal-living-memory-library

What This Skill Does

The LYGO Universal Living Memory Library (v1.1) is a specialized architectural framework designed to maintain long-term coherence and integrity within OpenClaw agents. Unlike standard memory management, which can become bloated and noisy, this skill enforces a strict 'Max 20' index policy. By limiting the active working set to twenty essential files, it ensures the agent maintains a high-signal, low-latency cognitive state. The skill introduces the {FRAGILE} tagging system, which allows users to explicitly mark memory sectors that require future manual review or reconciliation. Through the LYGO-MINT protocol, the library provides cryptographic provenance, enabling agents to verify the authenticity and drift-status of their internal records via hashes and anchors. This is a purely advisory skill; it acts as a gatekeeper and curator for the agent’s memory without autonomous executive control.

Installation

To integrate this library into your OpenClaw environment, ensure you have access to the clawhub repository. Execute the following command in your terminal or agent interface:

clawhub install openclaw/skills/skills/deepseekoracle/lygo-universal-living-memory-library

Once installed, you must verify the environment by running the included scripts/self_check.py to ensure all path mappings and reference files (library_spec.md, audit_protocol.md) are correctly identified by your system.

Use Cases

This skill is ideal for complex, multi-stage projects where drift is a significant risk. It is perfect for long-running research tasks where you need to preserve specific core insights across agent resets or context shifts. Developers use this to maintain a 'Master Archive' of documentation, while project managers use the {FRAGILE} tagging to highlight incomplete or high-variance data sets. If your agent is tasked with summarizing massive volumes of logs, the compression workflow distills this input into a clean, audit-ready index.

Example Prompts

1. "Run Living Memory Audit (max20 index) and report any files currently tagged as {FRAGILE} or showing significant drift."
2. "Compress these project logs into a Master Archive using the defined Living Memory protocols."
3. "Mint the current Master Archive with LYGO-MINT and provide the resulting Anchor Snippet for project verification."

Tips & Limitations

• The 'Max 20' constraint is hard-coded for efficiency; do not attempt to bypass this as it may degrade agent reasoning performance due to context congestion.
• Always perform a manual review of {FRAGILE} tags before running the full compression workflow, as compressed archives may prune transient data that is not explicitly indexed.
• Ensure that your system allows file-read/write access to the references/ directory to allow the scripts to update the metadata hashes correctly.