skillpp 0.1.0

package/COMPATIBILITY.md

@@ -0,0 +1,58 @@
+# Skill++ Compatibility Contract
+
+Skill++ treats routing, schemas, checkpoints, and CLI commands as a public protocol. New skills can be added, but existing v0.1 behavior must remain available unless the package opens a new compatibility baseline.
+
+## Public Contract
+
+The following surfaces are stable within the v0.1 line:
+
+- npm package name: `skillpp`
+- CLI command: `skillpp`
+- AI entry point: `SKILL.md`
+- machine registry: `skillpp.manifest.json`
+- schema files: `handoff`, `token`, `audit`, and `checkpoint`
+- pipeline IDs beginning with `P_`
+- checkpoint IDs and blocking semantics
+- command routes for `analyze`, `scan`, `trade`, `audit`, `wallet`, `signals`, and `create`
+- default private-path redaction in `skillpp doctor`
+
+## Update Rules
+
+Use additive changes for v0.1:
+
+- Add a new skill without renaming or removing existing skill names.
+- Add a new pipeline without changing existing pipeline IDs.
+- Add fields to schemas without removing required fields.
+- Add optional commands without changing existing command meanings.
+- Add richer adapters and prompts while keeping `SKILL.md` as the primary AI entry.
+
+Treat these as breaking changes:
+
+- Removing or renaming an existing skill, pipeline ID, checkpoint ID, schema file, or CLI command.
+- Reordering a pipeline in a way that changes the required v0.1 sequence.
+- Changing a blocking checkpoint into a non-blocking checkpoint.
+- Exposing absolute local paths in default `skillpp doctor` output.
+- Making write operations run without explicit checkpoint confirmation.
+
+Breaking changes require a new compatibility baseline. After Skill++ reaches `1.0.0`, breaking protocol changes also require a new major package version.
+
+## Compatibility Tests
+
+The v0.1 baseline is stored in `tests/compatibility/v0.1.0.json`.
+
+Run:
+
+```bash
+npm run compatibility
+```
+
+The check verifies that:
+
+- all required v0.1 skills still exist
+- all required v0.1 pipeline IDs still exist
+- existing pipeline sequences remain in order
+- blocking checkpoint behavior remains blocking
+- CLI commands still route to the same pipeline IDs
+- package metadata still exposes the same npm command and core files
+
+The baseline is intentionally separate from implementation details. It allows Skill++ to keep growing while protecting the protocol users already rely on.

package/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Skill++
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

package/README.md

@@ -0,0 +1,307 @@
+# Skill++
+
+**A multi-agent skill orchestration protocol for on-chain intelligence.**
+
+![Skill++ banner](assets/skillpp-banner.png)
+
+Skill++ turns isolated AI skills into coordinated workflows for token research, wallet review, smart-contract audit, smart-money tracking, chain scanning, and guarded write-operation handoff.
+
+[简体中文](README.zh-CN.md) | [Website](https://skillpp.ai) | [X](https://x.com/SkillppAI) | [GitHub](https://github.com/skillpp/skillpp)
+
+![npm](https://img.shields.io/npm/v/skillpp?style=flat-square)
+![install](https://img.shields.io/badge/install-npm%20i%20-g%20skillpp-CB3837?style=flat-square)
+![Node](https://img.shields.io/badge/node-%3E%3D18-339933?style=flat-square)
+![License](https://img.shields.io/badge/license-MIT-111827?style=flat-square)
+![BinanceAI](https://img.shields.io/badge/BinanceAI-compatible-0B0E11?style=flat-square)
+![Claude Opus](https://img.shields.io/badge/Claude%20Opus-compatible-6B4EFF?style=flat-square)
+![GPT](https://img.shields.io/badge/GPT-compatible-111827?style=flat-square)
+![Gemini](https://img.shields.io/badge/Gemini-compatible-1A73E8?style=flat-square)
+![Kimi](https://img.shields.io/badge/Kimi-compatible-111827?style=flat-square)
+![Mimo](https://img.shields.io/badge/Mimo-compatible-16A34A?style=flat-square)
+![OpenClaw](https://img.shields.io/badge/OpenClaw-compatible-DC2626?style=flat-square)
+![Codex](https://img.shields.io/badge/Codex-compatible-000000?style=flat-square)
+
+## Overview
+
+Current status: v0.1 public npm release.
+
+AI skills are useful on their own, but production-grade on-chain analysis needs routing, sequencing, cross-checking, safety gates, and structured handoff. Skill++ provides that coordination layer.
+
+Skill++ is a protocol-first AI skill package. Some skills run local read-only CLIs; some call public read-only APIs; others emit structured handoff for agent-side execution. It is not a wallet, exchange, or Four.meme SDK, and it must not report a write operation as complete unless the required external tool actually ran.
+
+| Layer | What Skill++ Provides |
+|---|---|
+| Router | Parses URLs, addresses, symbols, source snippets, wallets, and natural-language intents |
+| Pipeline Protocol | Maps requests to 7 stable pipeline IDs |
+| Skill Registry | Registers 25 skills across Binance Web3, Binance Exchange, Four.meme, audit-plus, and Skill++ modules |
+| Handoff | Emits structured JSON so downstream skills can continue without losing context |
+| Risk Fusion | Cross-checks quick audit, deep audit, market data, smart-money signals, and wallet exposure |
+| Safety | Blocks sensitive actions behind checkpoints and explicit user confirmation |
+| Agent Adapters | Provides loading guidance for BinanceAI, Claude, GPT, Gemini, Kimi, Mimo, OpenClaw, and Codex |
+
+Skill++ is designed for both expert agents and users who do not know Web3 terminology. The agent can explain risk in plain language while still preserving machine-readable execution state.
+
+## Install
+
+Install from npm:
+
+```bash
+npm install -g skillpp
+skillpp doctor
+skillpp skills
+```
+
+GitHub source install:
+
+```bash
+npm install -g github:skillpp/skillpp
+skillpp doctor
+skillpp skills
+```
+
+Local repository use:
+
+```bash
+git clone https://github.com/skillpp/skillpp.git
+cd skillpp
+npm install -g .
+skillpp doctor
+```
+
+## Use With AI Agents
+
+Skill++ is a skill package, not only a CLI. After installation or cloning, point the AI agent at the package root and ask it to read `SKILL.md` first. If an agent only reads this README, it has seen the product overview but has not loaded the Skill++ protocol.
+
+Minimum loading prompt:
+
+```text
+Use Skill++ from this repository. Read SKILL.md first, then skillpp.manifest.json and the matching adapter under adapters/ for your runtime. Use scripts/skillpp.mjs --dry-run when deterministic routing is available.
+```
+
+The npm package includes `SKILL.md`, `skillpp.manifest.json`, `schemas/`, `adapters/`, `prompts/`, `skills/`, and the `skillpp` CLI, so compatible agents can learn the protocol directly from the installed package or the GitHub repository.
+
+## Quick Start
+
+```bash
+# Token analysis: metadata -> quick audit -> smart-money signal -> risk fusion
+skillpp analyze "0x55d398326f99059ff775485246999027b3197955" --dry-run
+
+# Chain scan: launchpad signals -> market rankings -> risk filtering -> opportunity board
+skillpp scan 56 --dry-run
+
+# Pre-trade safety check: audit -> risk fusion -> confirmation checkpoint
+skillpp trade "0x55d398326f99059ff775485246999027b3197955" --dry-run
+
+# Four.meme creation flow: safety notice -> creation checkpoint
+skillpp create "create a meme token on four.meme" --dry-run
+```
+
+Minimal expected output shape:
+
+```text
+$ skillpp parse "0x55d398326f99059ff775485246999027b3197955"
+{
+  "command": "parse",
+  "parsed": {
+    "type": "address",
+    "contractAddress": "0x55d398326f99059ff775485246999027b3197955"
+  }
+}
+
+$ skillpp analyze "0x55d398326f99059ff775485246999027b3197955" --dry-run
+Pipeline: P_TOKEN_ANALYSIS
+Steps: query-token-info -> query-token-audit -> trading-signal -> risk-fusion
+Checkpoint: AUDIT_RESULT
+Handoff: { "_meta": { "pipeline": "P_TOKEN_ANALYSIS" }, "results": { ... } }
+```
+
+## Privacy-Safe Doctor
+
+`skillpp doctor` is safe to paste into public issues by default. It reports package health without exposing absolute local paths.
+
+```bash
+skillpp doctor
+```
+
+Default output redacts local paths:
+
+```json
+{
+  "command": "doctor",
+  "packageRoot": "<redacted>",
+  "checks": {
+    "skillsDir": { "path": "skills", "exists": true }
+  }
+}
+```
+
+Use absolute paths only in a private local session:
+
+```bash
+skillpp doctor --show-paths
+```
+
+Do not paste `--show-paths` output into public GitHub issues, chat logs, or bug reports.
+
+## Naming
+
+The formal name is **Skill Plus Plus (Skill++)**.
+
+| Name | Use |
+|---|---|
+| Skill++ | Product name and documentation name |
+| Skill Plus Plus | Plain-English full name |
+| `skillpp` | npm package name and CLI command |
+| `skillpp/` | Recommended repository or local folder name |
+
+## Architecture
+
+```text
+User input
+  URL / address / symbol / source code / wallet / natural language
+        |
+        v
+Skill++ scheduler
+  1. Parse input type, chain, and address
+  2. Check tools and dependencies
+  3. Match a stable pipeline ID
+  4. Run CLI, API, or text-based AI skills
+  5. Trigger safety checkpoints
+  6. Emit handoff JSON
+  7. Recommend next routes
+        |
+        v
+25 coordinated skills
+  Binance Web3 / Binance Exchange / Four.meme / audit-plus / Skill++ Library
+```
+
+## Pipelines
+
+| ID | Purpose | Sequence |
+|---|---|---|
+| `P_TOKEN_ANALYSIS` | Token analysis | token info -> token audit -> smart-money signal -> risk fusion |
+| `P_CHAIN_SCAN` | Chain opportunity scan | meme rush -> market rank -> risk fusion -> opportunity board |
+| `P_TRADE_SAFETY` | Pre-trade safety | token audit -> audit-plus -> risk fusion |
+| `P_WALLET_XRAY` | Wallet analysis | wallet positions -> per-token risk -> wallet-doctor |
+| `P_SMART_MONEY` | Smart-money tracking | trading signal -> token risk -> risk fusion |
+| `P_FOURMEME_CREATE` | Four.meme creation | integration flow -> safety checkpoint |
+| `P_DEEP_AUDIT` | Deep contract audit | token info -> token audit -> contract-profiler -> audit-plus -> risk fusion |
+
+Pipeline IDs are stable protocol identifiers. Skill++ does not rely on drifting numbered pipeline labels.
+
+## Skill Inventory
+
+| Group | Skills |
+|---|---|
+| Binance Web3 | `query-token-info`, `query-token-audit`, `query-address-info`, `trading-signal`, `crypto-market-rank`, `meme-rush`, `binance-agentic-wallet`, `binance-tokenized-securities-info` |
+| Binance Exchange and Payments | `binance`, `fiat`, `p2p`, `payment`, `square-post`, `onchain-pay` |
+| Four.meme | `four-meme-integration`, `four-meme-ai`, `four-guard` |
+| Audit Core | `audit-plus` |
+| Skill++ Library | `contract-profiler`, `risk-fusion`, `wallet-doctor`, `newbie-tutor`, `watchtower`, `opportunity-board`, `scam-pattern-lab` |
+
+## Compatibility
+
+Skill++ treats pipeline IDs, command routes, schemas, checkpoints, and default safety behavior as a public protocol. The v0.1 compatibility contract is documented in [COMPATIBILITY.md](COMPATIBILITY.md) and enforced by CI.
+
+For v0.1 updates, add new skills, fields, prompts, adapters, or pipelines without renaming or removing existing public identifiers. Breaking changes require a new compatibility baseline; after `1.0.0`, they also require a new major version.
+
+## Skill++ Library Modules
+
+| Module | Role |
+|---|---|
+| `contract-profiler` | Profiles contract structure before deep audit |
+| `risk-fusion` | Combines token audit, audit-plus, market, smart-money, wallet, and social signals |
+| `wallet-doctor` | Diagnoses wallet portfolio health and exposure |
+| `newbie-tutor` | Explains on-chain and audit terms using current context only |
+| `watchtower` | Designs monitoring rules for addresses, contracts, and events |
+| `opportunity-board` | Aggregates scan, ranking, smart-money, and risk-filtered candidates |
+| `scam-pattern-lab` | Maps audit evidence to known on-chain scam patterns |
+
+## AI Agent Compatibility
+
+Skill++ uses `SKILL.md` as the primary AI entry point. Adapters and prompts are included for agent-specific loading.
+
+| AI / Agent | Loading Path |
+|---|---|
+| BinanceAI | Read `SKILL.md`, then use `adapters/binance-ai.md` |
+| Claude / Claude Code / Claude Opus | Read `SKILL.md`, then use `adapters/claude.md` |
+| GPT / ChatGPT / Custom GPT | Use `adapters/gpt.md` or `prompts/universal-system-prompt.md` |
+| Gemini | Use `adapters/gemini.md` |
+| Kimi | Use `adapters/kimi.md` |
+| Mimo | Use `adapters/mimo.md` |
+| OpenClaw | Use `adapters/openclaw.md` |
+| Codex | Read `SKILL.md`; run `scripts/skillpp.mjs` for deterministic routing |
+
+## Write-Operation Boundary
+
+The `skillpp` package includes the scheduler, protocol files, schemas, adapters, prompts, and bundled read-only helper scripts. It does not silently install wallet, exchange, or Four.meme write-operation tools.
+
+Write-operation flows are checkpointed handoffs unless the user installs and explicitly invokes the matching external CLI.
+
+Install external write-operation CLIs only when needed:
+
+```bash
+npm i -g @binance/agentic-wallet
+npm i -g @binance/binance-cli
+npm i -g @four-meme/four-meme-ai@latest
+```
+
+These provide `baw`, `binance-cli`, and `fourmeme`.
+
+Without these tools, Skill++ can still parse, route, run read-only checks, audit, fuse risk signals, and emit checkpoints. It must not claim that a write operation succeeded when the required external CLI is unavailable.
+
+## Safety Model
+
+Skill++ can execute read-only Node CLIs for token lookup, rankings, smart-money data, and wallet positions. It does not silently execute operations involving:
+
+- private keys, API keys, login state, transfers, swaps, orders, publishing, or posting
+- high risk, conflicting results, or insufficient evidence
+- any action where the user carries funds or identity risk
+
+Sensitive actions produce structured checkpoints and require explicit user confirmation. Blocking checkpoints exit with code `10` in `scripts/skillpp.mjs`.
+
+Skill++ output is for research and risk review only. It is not financial advice.
+
+## Repository Layout
+
+```text
+skillpp/
+├── README.md
+├── README.zh-CN.md
+├── COMPATIBILITY.md
+├── package.json
+├── SKILL.md
+├── registry.md
+├── pipelines.md
+├── rules.md
+├── skillpp.manifest.json
+├── schemas/
+├── scripts/
+├── adapters/
+├── prompts/
+├── assets/
+├── tests/
+└── skills/
+    ├── binance-web3/
+    ├── binance/
+    ├── four-meme/
+    ├── audit-plus/
+    └── skillpp/
+```
+
+Skill paths use `skills/<group>/<skill-name>/`.
+
+## Release Checks
+
+```bash
+npm test
+npm run validate
+npm run compatibility
+npm pack --dry-run
+```
+
+GitHub Actions runs the same checks on push, pull request, and manual dispatch.
+
+## License
+
+[MIT](LICENSE)

package/README.zh-CN.md

@@ -0,0 +1,307 @@
+# Skill++
+
+**面向链上情报的多智能体 Skill 调度协议。**
+
+![Skill++ banner](assets/skillpp-banner.png)
+
+Skill++ 把孤立的 AI skills 组织成可调度的工作流，用于代币研究、钱包体检、合约审计、聪明钱追踪、扫链分析和带安全检查点的写操作交接。
+
+[English](README.md) | [官网](https://skillpp.ai) | [X](https://x.com/SkillppAI) | [GitHub](https://github.com/skillpp/skillpp)
+
+![npm](https://img.shields.io/npm/v/skillpp?style=flat-square)
+![install](https://img.shields.io/badge/install-npm%20i%20-g%20skillpp-CB3837?style=flat-square)
+![Node](https://img.shields.io/badge/node-%3E%3D18-339933?style=flat-square)
+![License](https://img.shields.io/badge/license-MIT-111827?style=flat-square)
+![BinanceAI](https://img.shields.io/badge/BinanceAI-compatible-0B0E11?style=flat-square)
+![Claude Opus](https://img.shields.io/badge/Claude%20Opus-compatible-6B4EFF?style=flat-square)
+![GPT](https://img.shields.io/badge/GPT-compatible-111827?style=flat-square)
+![Gemini](https://img.shields.io/badge/Gemini-compatible-1A73E8?style=flat-square)
+![Kimi](https://img.shields.io/badge/Kimi-compatible-111827?style=flat-square)
+![Mimo](https://img.shields.io/badge/Mimo-compatible-16A34A?style=flat-square)
+![OpenClaw](https://img.shields.io/badge/OpenClaw-compatible-DC2626?style=flat-square)
+![Codex](https://img.shields.io/badge/Codex-compatible-000000?style=flat-square)
+
+## 概览
+
+当前状态：v0.1 公开 npm 发布版。
+
+单个 AI skill 可以解决局部问题，但真正的链上分析需要路由、排序、交叉验证、安全闸门和结构化交接。Skill++ 提供这层调度协议。
+
+Skill++ 是协议优先的 AI skill 包。部分 skill 会运行本地只读 CLI，部分会调用公开只读 API，另一些会输出结构化 handoff 交给智能体继续执行。它不是钱包、交易所或 Four.meme SDK；除非所需外部工具真实执行成功，否则不能声称写操作已完成。
+
+| 层级 | Skill++ 提供什么 |
+|---|---|
+| Router | 解析链接、地址、符号、源码片段、钱包和自然语言意图 |
+| Pipeline Protocol | 把请求映射到 7 条稳定流水线 ID |
+| Skill Registry | 注册 Binance Web3、Binance Exchange、Four.meme、audit-plus 和 Skill++ 模块共 25 个 skill |
+| Handoff | 输出结构化 JSON，让下游 skill 不丢上下文 |
+| Risk Fusion | 交叉验证快速审计、深度审计、市场数据、聪明钱信号和钱包风险 |
+| Safety | 对敏感动作强制检查点和用户确认 |
+| Agent Adapters | 适配 BinanceAI、Claude、GPT、Gemini、Kimi、Mimo、OpenClaw 和 Codex |
+
+Skill++ 同时面向专业智能体和不熟悉 Web3 的用户。它可以用人话解释风险，也保留机器可读的执行状态。
+
+## 安装
+
+npm 安装：
+
+```bash
+npm install -g skillpp
+skillpp doctor
+skillpp skills
+```
+
+GitHub 源码安装：
+
+```bash
+npm install -g github:skillpp/skillpp
+skillpp doctor
+skillpp skills
+```
+
+本地仓库使用：
+
+```bash
+git clone https://github.com/skillpp/skillpp.git
+cd skillpp
+npm install -g .
+skillpp doctor
+```
+
+## 给 AI 智能体使用
+
+Skill++ 不只是 CLI，也是一个 skill 包。安装或克隆后，把包根目录交给 AI 智能体，并要求它先读取 `SKILL.md`。如果智能体只读 README，它只看到了项目概览，还没有真正加载 Skill++ 协议。
+
+最小加载提示词：
+
+```text
+Use Skill++ from this repository. Read SKILL.md first, then skillpp.manifest.json and the matching adapter under adapters/ for your runtime. Use scripts/skillpp.mjs --dry-run when deterministic routing is available.
+```
+
+npm 包会包含 `SKILL.md`、`skillpp.manifest.json`、`schemas/`、`adapters/`、`prompts/`、`skills/` 和 `skillpp` CLI，因此兼容的智能体可以直接从已安装包或 GitHub 仓库学习协议。
+
+## 快速开始
+
+```bash
+# 代币分析：基础信息 -> 快速审计 -> 聪明钱信号 -> 风险融合
+skillpp analyze "0x55d398326f99059ff775485246999027b3197955" --dry-run
+
+# 扫链：发射台信号 -> 市场榜单 -> 风险过滤 -> 机会看板
+skillpp scan 56 --dry-run
+
+# 交易前安全检查：审计 -> 风险融合 -> 确认检查点
+skillpp trade "0x55d398326f99059ff775485246999027b3197955" --dry-run
+
+# Four.meme 创建流程：安全提示 -> 创建检查点
+skillpp create "create a meme token on four.meme" --dry-run
+```
+
+最小输出形态：
+
+```text
+$ skillpp parse "0x55d398326f99059ff775485246999027b3197955"
+{
+  "command": "parse",
+  "parsed": {
+    "type": "address",
+    "contractAddress": "0x55d398326f99059ff775485246999027b3197955"
+  }
+}
+
+$ skillpp analyze "0x55d398326f99059ff775485246999027b3197955" --dry-run
+Pipeline: P_TOKEN_ANALYSIS
+Steps: query-token-info -> query-token-audit -> trading-signal -> risk-fusion
+Checkpoint: AUDIT_RESULT
+Handoff: { "_meta": { "pipeline": "P_TOKEN_ANALYSIS" }, "results": { ... } }
+```
+
+## 隐私安全自检
+
+`skillpp doctor` 默认可以安全粘贴到公开 issue。它会报告包健康状态，但不会暴露本机绝对路径。
+
+```bash
+skillpp doctor
+```
+
+默认输出会隐藏本机路径：
+
+```json
+{
+  "command": "doctor",
+  "packageRoot": "<redacted>",
+  "checks": {
+    "skillsDir": { "path": "skills", "exists": true }
+  }
+}
+```
+
+只有在本地私有会话中才使用：
+
+```bash
+skillpp doctor --show-paths
+```
+
+不要把 `--show-paths` 输出粘贴到公开 GitHub issue、聊天记录或问题报告。
+
+## 命名
+
+正式名称是 **Skill Plus Plus（Skill++）**。
+
+| 名称 | 用途 |
+|---|---|
+| Skill++ | 产品名和文档名 |
+| Skill Plus Plus | 不使用符号时的英文全称 |
+| `skillpp` | npm 包名和 CLI 命令 |
+| `skillpp/` | 推荐的仓库名或本地目录名 |
+
+## 架构
+
+```text
+用户输入
+  链接 / 地址 / 符号 / 源码 / 钱包 / 自然语言
+        |
+        v
+Skill++ 调度器
+  1. 解析输入类型、链和地址
+  2. 检查工具和依赖
+  3. 匹配稳定流水线 ID
+  4. 执行 CLI、API 或文本型 AI skill
+  5. 触发安全检查点
+  6. 输出 handoff JSON
+  7. 推荐下一步路由
+        |
+        v
+25 个 skill 协同工作
+  Binance Web3 / Binance Exchange / Four.meme / audit-plus / Skill++ Library
+```
+
+## 流水线
+
+| ID | 用途 | 顺序 |
+|---|---|---|
+| `P_TOKEN_ANALYSIS` | 代币分析 | token info -> token audit -> smart-money signal -> risk fusion |
+| `P_CHAIN_SCAN` | 扫链找机会 | meme rush -> market rank -> risk fusion -> opportunity board |
+| `P_TRADE_SAFETY` | 交易前安全 | token audit -> audit-plus -> risk fusion |
+| `P_WALLET_XRAY` | 钱包分析 | wallet positions -> per-token risk -> wallet-doctor |
+| `P_SMART_MONEY` | 聪明钱追踪 | trading signal -> token risk -> risk fusion |
+| `P_FOURMEME_CREATE` | Four.meme 创建 | integration flow -> safety checkpoint |
+| `P_DEEP_AUDIT` | 合约深度审计 | token info -> token audit -> contract-profiler -> audit-plus -> risk fusion |
+
+流水线 ID 是稳定协议标识，不依赖容易漂移的数字编号。
+
+## Skill 清单
+
+| 分组 | Skills |
+|---|---|
+| Binance Web3 | `query-token-info`, `query-token-audit`, `query-address-info`, `trading-signal`, `crypto-market-rank`, `meme-rush`, `binance-agentic-wallet`, `binance-tokenized-securities-info` |
+| Binance Exchange and Payments | `binance`, `fiat`, `p2p`, `payment`, `square-post`, `onchain-pay` |
+| Four.meme | `four-meme-integration`, `four-meme-ai`, `four-guard` |
+| Audit Core | `audit-plus` |
+| Skill++ Library | `contract-profiler`, `risk-fusion`, `wallet-doctor`, `newbie-tutor`, `watchtower`, `opportunity-board`, `scam-pattern-lab` |
+
+## 兼容性
+
+Skill++ 把 pipeline ID、命令路由、schema、checkpoint 和默认安全行为都视为公开协议。v0.1 兼容性合约写在 [COMPATIBILITY.md](COMPATIBILITY.md)，并由 CI 自动检查。
+
+v0.1 更新应采用增量方式：可以添加新 skill、新字段、新 prompt、新 adapter 或新 pipeline，但不能重命名或删除已有公开标识。破坏性变更需要新的兼容性基线；`1.0.0` 之后还需要新的 major 版本。
+
+## Skill++ 库模块
+
+| 模块 | 作用 |
+|---|---|
+| `contract-profiler` | 深度审计前提取合约结构 |
+| `risk-fusion` | 融合 token audit、audit-plus、市场、聪明钱、钱包和社交信号 |
+| `wallet-doctor` | 诊断钱包持仓健康度和风险暴露 |
+| `newbie-tutor` | 只基于当前上下文解释链上和审计术语 |
+| `watchtower` | 为地址、合约和事件设计监控规则 |
+| `opportunity-board` | 聚合扫链、榜单、聪明钱和风险过滤候选 |
+| `scam-pattern-lab` | 把审计证据映射到常见链上骗局模式 |
+
+## AI 兼容性
+
+Skill++ 以 `SKILL.md` 作为主要 AI 入口，并提供不同 agent 的 adapter 和 prompt。
+
+| AI / Agent | 加载方式 |
+|---|---|
+| BinanceAI | 读取 `SKILL.md`，再使用 `adapters/binance-ai.md` |
+| Claude / Claude Code / Claude Opus | 读取 `SKILL.md`，再使用 `adapters/claude.md` |
+| GPT / ChatGPT / Custom GPT | 使用 `adapters/gpt.md` 或 `prompts/universal-system-prompt.md` |
+| Gemini | 使用 `adapters/gemini.md` |
+| Kimi | 使用 `adapters/kimi.md` |
+| Mimo | 使用 `adapters/mimo.md` |
+| OpenClaw | 使用 `adapters/openclaw.md` |
+| Codex | 读取 `SKILL.md`；需要确定性路由时运行 `scripts/skillpp.mjs` |
+
+## 写操作边界
+
+`skillpp` 包包含调度器、协议文件、schemas、adapters、prompts 和随包只读辅助脚本。它不会静默安装钱包、交易所或 Four.meme 写操作工具。
+
+写操作流程默认是带检查点的 handoff；只有用户安装并明确调用匹配的外部 CLI 后，才进入真实写操作执行。
+
+只有需要写操作时才安装外部 CLI：
+
+```bash
+npm i -g @binance/agentic-wallet
+npm i -g @binance/binance-cli
+npm i -g @four-meme/four-meme-ai@latest
+```
+
+这些包提供 `baw`、`binance-cli` 和 `fourmeme`。
+
+没有这些工具时，Skill++ 仍可解析、路由、执行只读检查、审计、风险融合和检查点输出，但不能声称写操作已经成功。
+
+## 安全模型
+
+Skill++ 可以直接执行部分只读 Node CLI，例如 token 查询、榜单、聪明钱数据和钱包持仓。它不会静默执行以下操作：
+
+- 私钥、API key、登录态、转账、swap、下单、发布或发文
+- 高风险、结果冲突或证据不足
+- 任何需要用户承担资金或身份风险的动作
+
+敏感动作会输出结构化检查点，并要求用户明确确认。`scripts/skillpp.mjs` 对 blocking checkpoint 使用 exit code `10`。
+
+Skill++ 输出仅用于研究和风险提示，不构成投资建议。
+
+## 目录结构
+
+```text
+skillpp/
+├── README.md
+├── README.zh-CN.md
+├── COMPATIBILITY.md
+├── package.json
+├── SKILL.md
+├── registry.md
+├── pipelines.md
+├── rules.md
+├── skillpp.manifest.json
+├── schemas/
+├── scripts/
+├── adapters/
+├── prompts/
+├── assets/
+├── tests/
+└── skills/
+    ├── binance-web3/
+    ├── binance/
+    ├── four-meme/
+    ├── audit-plus/
+    └── skillpp/
+```
+
+Skill 路径统一使用 `skills/<group>/<skill-name>/`。
+
+## 发布检查
+
+```bash
+npm test
+npm run validate
+npm run compatibility
+npm pack --dry-run
+```
+
+GitHub Actions 会在 push、pull request 和手动触发时运行同样检查。
+
+## License
+
+[MIT](LICENSE)