LYGO-MINT Verifier

将 Champion/prompt/workflow pack 规范化并计算确定性 SHA-256，在 --i-consent 同意下写入追加式与 canonical 账本，并生成可移植 Anchor Snippet 供人工发布。纯本地 Python 标准库，声明无网络、无 subprocess、无 shell，skill 自身不发布内容。

deepseekoracle

热度 65.95

最近更新 2026-08-21 11:55:45

分类：效率依赖：代码执行 / 读写文件应用场景：数据处理 / 工作流 / 模板

安装技能

复制下面的话术发给你的 Agent，它会通过 secure-skills 检索并安装该技能。

执行命令安装此 skill： npx secure-skills-cli add lygo-mint-verifier

Metadata

name: lygo-mint-verifier
description: LYGO-MINT Verifier — canonicalize a pack, deterministic SHA-256, append-only + canonical ledgers, portable Anchor Snippet for posting anywhere (Moltbook/X/Discord/4claw). Pure local stdlib: no subprocess, no network, no auto-publish. Ledger writes need --i-consent. Pairs with lygo-mint-walkthrough, lygo-continuum-integrator, lygo-geodesic-sealer. Use for Champion/alignment prompt packs and verifiable receipts.
version: 1.1.1
license: MIT-0
metadata: {"openclaw":{"emoji":"🔏","homepage":"https://github.com/DeepSeekOracle/lygo-protocol-stack/tree/main/clawhub/mirrors/lygo-mint-verifier","requires":{"anyBins":["python","python3"]}},"lygo":true,"mint":true,"signature":"Delta9Phi963-MINT-VERIFIER-v1.1.1","publisher":"deepseekoracle","clawhub":"https://clawhub.ai/deepseekoracle/lygo-mint-verifier","security_audit":"https://clawhub.ai/deepseekoracle/skills/lygo-mint-verifier/security-audit","github":"https://github.com/DeepSeekOracle/lygo-protocol-stack","security_review":"1.1.1-skillspector-consent-wrapper-honest","permissions":{"network":false,"shell":false,"subprocess":false,"filesystem":{"read":"operator --pack path","write":"skill state/ ledgers only with --i-consent"},"publish":false}}

LYGO-MINT Verifier v1.1.1

Turns an aligned Champion / prompt / workflow pack into a verifiable artifact:

canonical form (deterministic)
SHA-256 hash
append-only + canonical ledgers (consent-gated)
portable Anchor Snippet (you post it — skill never posts)

Signature: Delta9Phi963-MINT-VERIFIER-v1.1.1
Security audit: https://clawhub.ai/deepseekoracle/skills/lygo-mint-verifier/security-audit

What changed in 1.1.1

ClawHub security-audit Medium findings (Intent-Code Divergence + Description-Behavior Mismatch):

backfill_anchors.py no longer auto-appends --i-consent — pass-through only
Compat wrappers stay honest with the advertised consent gate
self_check asserts consent_wrapper_honest + backfill_requires_consent

What changed in 1.1.0

Removed subprocess — mint fully in-process
Declared permissions in SKILL.md / claw.json
Ledgers default under skill state/

Install

npx clawhub@latest install deepseekoracle/lygo-mint-verifier

Commands

cd path/to/lygo-mint-verifier
python scripts/self_check.py

# Mint (dry-run hash) / write ledgers with --i-consent
python scripts/mint_cli.py mint --pack ./pack.md --version 2026-08-20.v1
python scripts/mint_cli.py mint --pack ./pack.md --version 2026-08-20.v1 \
  --champion LYRA --anchor SEAL_55_T --i-consent --json

# Verify
python scripts/mint_cli.py verify --pack ./pack.md --hash <64-hex>

# Snippet / backfill
python scripts/mint_cli.py snippet --hash <64-hex> --title "My Pack"
python scripts/mint_cli.py backfill --hash <64-hex> --channel x --id https://x.com/... --i-consent

Compat wrappers (same flags, no subprocess, never inject --i-consent):
mint_pack_local.py · make_anchor_snippet.py · backfill_anchors.py

Command	Network	Subprocess	Writes
mint / verify / snippet	none	none	mint/backfill only with `--i-consent`

Pair with

Skill	Role
`lygo-mint-walkthrough`	Interactive tutorial
`lygo-continuum-integrator`	∫(Truth×Light) + phase-lock receipts
`lygo-geodesic-sealer`	\|ψ⟩ geodesic dual-ledger attest
`lygo-continuum`	Falsifiable work capsules

Security

Read references/SECURITY.md + references/SKILLSPECTOR_AUDIT.md.

No network · no subprocess · no shell
Never put API keys in packs
You remain the publisher

Δ9Φ963 — hash over vibes · consent before ledger · human posts the snippet.