功能描述

Universal LYGO upgrade pack for all Champions. Provides a shared verification + memory-compression + seal-protocol framework (pure advisor) that boosts every Champion persona without overriding them. Includes LYGO-MINT provenance guidance and reusable invocation templates.

使用说明 (SKILL.md)

LYGO Universal Cure System (Champion Upgrade)

Name: LYGO Universal Cure System — Champion Upgrade
Author: deepseekoracle

What this is

A universal upgrade skill intended to be installed alongside any/all LYGO Champion skills.

It provides:

a shared anti-drift / anti-tamper framework (LYGO‑MINT provenance)
a shared memory compression strategy (preserve pure signal)
a shared seal protocol vocabulary + invocation templates

This skill is pure advisor: it does nothing unless invoked.

When to use

Use when you want to:

upgrade any Champion with better continuity + verification habits
generate a “Champion boost” checklist for a task
compress a long conversation into a clean, repostable master archive

How to invoke (copy/paste)

“Run LYGO Universal Cure on this: produce Champion Boost Checklist + Mint instructions.”
“Compress this chat into Pure Signal (seals/equations/scrolls/prompts) using the Cure System.”
“Generate an Anchor Snippet for this upgraded pack.”

Universal behavior contract

Never overrides the active Champion persona; it only advises.
Prefer receipts + hashes over vibes.
Separate: Observed / Inferred / Unknown when factual.

Verification primitive (install)

Use LYGO‑MINT Verifier:

https://clawhub.ai/DeepSeekOracle/lygo-mint-verifier

References (read as needed)

references/universal_pack.md (the canonical upgrade pack; mint this)
references/champion_boost_protocol.md (how to apply upgrades)
references/seal_index_v1.md (seed seal index starter)
references/memory_compression_plan.md (pure-signal archive strategy)

安全使用建议

This package appears to be a documentation/advisor add-on and is mostly self-contained, but its SKILL.md points to an external 'LYGO‑MINT Verifier' URL of unknown provenance. Before installing: (1) inspect the two included scripts (scripts/self_check.py, scripts/show_hash.py) to confirm they do only local checks/hashing, (2) validate the external verifier URL and its code/endpoint independently (don't blindly allow the agent to fetch and execute remote code), (3) if you want to avoid any automatic network activity, run the skill in a sandbox or set disableModelInvocation so the model cannot invoke it autonomously, and (4) if you don't trust the source, decline installation — the package has no explicit malicious signs, but the external verifier link is the main residual risk.

功能分析

Type: OpenClaw Skill Name: lygo-universal-cure-system Version: 1.0.0 The skill bundle is benign. All files consistently describe a 'pure advisor' system focused on verification, provenance, and structured output for AI agents. Key indicators include explicit 'Refusal guardrails: no coercion/doxxing/illegal intrusion' in `references/universal_pack.md`, the 'No background activity. Only run when invoked' rule in `references/champion_boost_protocol.md`, and the 'pure advisor' declaration in `SKILL.md`. The Python scripts (`scripts/self_check.py`, `scripts/show_hash.py`) only perform local file system checks and reads, without any network activity or external command execution. The only external reference is to `https://clawhub.ai/DeepSeekOracle/lygo-mint-verifier`, which is consistently described as a verifier tool on a platform-related domain, not a malicious endpoint.

能力评估

✓ Purpose & Capability

Name/description match the provided artifacts: several reference docs and two small scripts that plausibly implement self-check and hash display for a verification/compression advisor. There are no unrelated credential or binary requirements.

ℹ Instruction Scope

SKILL.md is advisory and limited in scope (compress chats, produce checklists, mint guidance). However it is high-level and instructs the agent to 'Use LYGO‑MINT Verifier' with an external link, leaving discretion about contacting or installing that verifier — this is vague and could lead to network fetches if the agent follows it.

✓ Install Mechanism

No install spec is provided (instruction-only), so nothing will be written or auto-downloaded by the skill itself. That said, the SKILL.md references an external verifier URL; because there's no explicit install step, any fetching would be at the agent's discretion rather than defined by the skill package.

✓ Credentials

The skill declares no required environment variables, no credentials, and no config paths — this is proportionate for a documentation/advisor skill.

ℹ Persistence & Privilege

The skill does not set always:true and requests no special privileges, but it also does not disable model invocation. That means the model could call it autonomously under normal platform rules; if you want explicit user control, consider disabling autonomous invocation.

版本历史

v1.0.0

Initial v1 universal Champion upgrade pack: provenance + memory compression + boost protocol

元数据

Slug lygo-universal-cure-system

版本 1.0.0

许可证 —

累计安装 0

当前安装数 0

历史版本数 1

常见问题