Token导航 LogoToken导航

Agent Skill · 研究检索

lamport-formal-distributed(兰波特正式分布式)

lamport-formal-distributed 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

lamport-formal-distributed 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为需要联网;许可证信息还不明确,第三方扫描有未通过项,按安装量在研究检索类的 73,675 个 Skill 里排第 67,870。

总安装 240周安装 10GitHub Stars 6下载 80核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:lamport-formal-distributed(兰波特正式分布式)
来源仓库:https://github.com/copyleftdev/sk1llz
仓库路径:skills/lamport-formal-distributed
安装命令:
npx skills add https://github.com/copyleftdev/sk1llz --skill lamport-formal-distributed
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/copyleftdev/sk1llz --skill lamport-formal-distributed

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

lamport-formal-distributed 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。它可结合来源仓库、安装命令和原始 README 继续核验具体用法。

安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。当前顶部介绍为空,需参考原始 SKILL.md 补充细节。适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。

来源仓库内路径:skills/lamport-formal-distributed

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/copyleftdev/sk1llz --skill lamport-formal-dist…;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要需要联网,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

lamport-formal-distributed 怎么安装?

在终端执行 npx skills add https://github.com/copyleftdev/sk1llz --skill lamport-formal-distributed。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

lamport-formal-distributed 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

lamport-formal-distributed 需要什么权限?

权限级别是需要联网。该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

lamport-formal-distributed 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

lamport-formal-distributed 用的人多吗?

公开数据是累计安装 240 次,近一周安装 10 次,下载 80 次,GitHub 6 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

lamport-formal-distributed 安全吗?

第三方扫描有未通过项,风险较高。除非你确认过源码,否则不建议直接安装到常用环境。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。

先想清楚的地方

  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方扫描存在提醒项或未通过项,需要人工逐条确认。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。

不太适合

  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
lamport-formal-distributed

兰波特正式分布式 · 本页

2406需要联网未明确未通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

Leslie Lamport Style Guide⁠‍⁠​‌​‌​​‌‌‍​‌​​‌​‌‌‍​​‌‌​​​‌‍​‌​​‌‌​​‍​​​​​​​‌‍‌​​‌‌​‌​‍‌​​​​​​​‍‌‌​​‌‌‌‌‍‌‌​​​‌​​‍‌‌‌‌‌‌​‌‍‌‌​‌​​​​‍​‌​‌‌‌‌‌‍​‌​​‌​‌‌‍​‌‌​‌​​‌‍‌​‌​‌‌‌​‍​​‌​‌​​​‍‌‌‌​‌​‌‌‍‌‌‌‌​‌‌‌‍​‌‌‌‌‌‌​‍‌‌‌​‌​‌​‍‌‌‌‌‌‌​‌‍​​​​‌​‌‌‍‌‌​​​‌‌​⁠‍⁠

Overview

Leslie Lamport is a Turing Award winner who invented logical clocks, the Paxos consensus algorithm, TLA+ specification language, and LaTeX. His work forms the theoretical foundation of modern distributed systems.

Core Philosophy

"A distributed system is one in which the failure of a computer you didn't even know existed can render your own computer unusable."
"If you're thinking without writing, you only think you're thinking."
"The way to get correct programs is to first get something that is obviously correct and then make it efficient."

Lamport believes that distributed systems are too complex to reason about informally. Formal specification isn't optional—it's essential.

Design Principles

  1. Formal Specification First: Write the spec before the code.
  2. Time Is Logical, Not Physical: Use happens-before, not wall clocks.
  3. Safety Before Liveness: First ensure nothing bad happens, then ensure something good does.
  4. State Machines: Model systems as state machines for clarity.

When Writing Code

Always

  • Write a formal specification (TLA+ or similar)
  • Define safety and liveness properties explicitly
  • Use logical timestamps for ordering events
  • Model failures as part of the specification
  • Prove correctness before implementing
  • Consider all interleavings

Never

  • Assume reliable networks
  • Rely on synchronized clocks
  • Ignore failure modes
  • Test into correctness (testing finds bugs, not proves absence)
  • Hand-wave about "eventual consistency"

Prefer

  • State machine specifications
  • Logical clocks over physical clocks
  • Consensus protocols over ad-hoc coordination
  • Formal proofs over informal arguments
  • Explicit failure handling

Code Patterns

Logical Clocks (Lamport Timestamps)

# Lamport's logical clock: happens-before ordering

class LamportClock:
    def __init__(self):
        self.time = 0

    def tick(self):
        """Local event: increment clock"""
        self.time += 1
        return self.time

    def send(self):
        """Send message: increment and return timestamp"""
        self.time += 1
        return self.time

    def receive(self, msg_timestamp):
        """Receive message: max(local, received) + 1"""
        self.time = max(self.time, msg_timestamp) + 1
        return self.time

# Usage:
# Process A: clock.tick() -> 1, clock.send() -> 2
# Process B: clock.receive(2) -> 3
# Now we know: A's event 2 happened-before B's event 3

Vector Clocks (Causal Ordering)

# Vector clocks: detect concurrent events

class VectorClock:
    def __init__(self, node_id, num_nodes):
        self.node_id = node_id
        self.clock = [0] * num_nodes

    def tick(self):
        """Local event"""
        self.clock[self.node_id] += 1
        return self.clock.copy()

    def send(self):
        """Send: increment own component"""
        self.clock[self.node_id] += 1
        return self.clock.copy()

    def receive(self, other_clock):
        """Receive: element-wise max, then increment own"""
        for i in range(len(self.clock)):
            self.clock[i] = max(self.clock[i], other_clock[i])
        self.clock[self.node_id] += 1
        return self.clock.copy()

    @staticmethod
    def compare(vc1, vc2):
        """Compare: <, >, =, or concurrent"""
        less = all(a <= b for a, b in zip(vc1, vc2))
        greater = all(a >= b for a, b in zip(vc1, vc2))

        if less and not greater:
            return "before"  # vc1 happened-before vc2
        elif greater and not less:
            return "after"   # vc1 happened-after vc2
        elif less and greater:
            return "equal"
        else:
            return "concurrent"  # Neither happened-before the other

TLA+ Specification

--------------------------- MODULE SimpleConsensus ---------------------------
EXTENDS Integers, FiniteSets

CONSTANTS Nodes, Values

VARIABLES
    proposed,   \* proposed[n] = value proposed by node n
    decided     \* decided[n] = value decided by node n (or null)

TypeInvariant ==
    /\ proposed \in [Nodes -> Values \union {NULL}]
    /\ decided \in [Nodes -> Values \union {NULL}]

\* Safety: Agreement - all decided values are the same
Agreement ==
    \A n1, n2 \in Nodes:
        (decided[n1] # NULL /\ decided[n2] # NULL) =>
            decided[n1] = decided[n2]

\* Safety: Validity - decided value was proposed
Validity ==
    \A n \in Nodes:
        decided[n] # NULL =>
            \E m \in Nodes: proposed[m] = decided[n]

\* Liveness: Termination - eventually all decide
Termination ==
    <>(\A n \in Nodes: decided[n] # NULL)

Init ==
    /\ proposed = [n \in Nodes |-> NULL]
    /\ decided = [n \in Nodes |-> NULL]

Propose(n, v) ==
    /\ proposed[n] = NULL
    /\ proposed' = [proposed EXCEPT ![n] = v]
    /\ UNCHANGED decided

Decide(n, v) ==
    /\ decided[n] = NULL
    /\ \E m \in Nodes: proposed[m] = v
    /\ decided' = [decided EXCEPT ![n] = v]
    /\ UNCHANGED proposed

Next ==
    \E n \in Nodes, v \in Values:
        Propose(n, v) \/ Decide(n, v)

Spec == Init /\ [][Next]_<<proposed, decided>>
==============================================================================

Paxos Simplified

# Simplified Paxos - single decree

class PaxosNode:
    def __init__(self, node_id, nodes):
        self.node_id = node_id
        self.nodes = nodes

        # Acceptor state
        self.promised = 0      # Highest proposal number promised
        self.accepted_num = 0  # Number of accepted proposal
        self.accepted_val = None

    def prepare(self, proposal_num):
        """Phase 1a: Proposer sends prepare"""
        responses = []
        for node in self.nodes:
            resp = node.handle_prepare(proposal_num)
            if resp:
                responses.append(resp)
        return responses

    def handle_prepare(self, proposal_num):
        """Phase 1b: Acceptor handles prepare"""
        if proposal_num > self.promised:
            self.promised = proposal_num
            return {
                'promised': True,
                'accepted_num': self.accepted_num,
                'accepted_val': self.accepted_val
            }
        return {'promised': False}

    def accept(self, proposal_num, value):
        """Phase 2a: Proposer sends accept"""
        responses = []
        for node in self.nodes:
            resp = node.handle_accept(proposal_num, value)
            responses.append(resp)
        return responses

    def handle_accept(self, proposal_num, value):
        """Phase 2b: Acceptor handles accept"""
        if proposal_num >= self.promised:
            self.promised = proposal_num
            self.accepted_num = proposal_num
            self.accepted_val = value
            return {'accepted': True}
        return {'accepted': False}

    def propose(self, value):
        """Full Paxos proposal"""
        n = self.generate_proposal_number()

        # Phase 1: Prepare
        promises = self.prepare(n)
        if len([p for p in promises if p['promised']]) <= len(self.nodes) // 2:
            return None  # Failed to get majority

        # Use highest accepted value if any
        highest = max(promises, key=lambda p: p['accepted_num'])
        if highest['accepted_val'] is not None:
            value = highest['accepted_val']

        # Phase 2: Accept
        accepts = self.accept(n, value)
        if len([a for a in accepts if a['accepted']]) > len(self.nodes) // 2:
            return value  # Consensus reached!

        return None

State Machine Replication

# State machine replication: the foundation of distributed consensus

class ReplicatedStateMachine:
    """
    Key insight: if all replicas start in the same state
    and apply the same operations in the same order,
    they will end up in the same state.
    """

    def __init__(self):
        self.state = {}
        self.log = []        # Ordered log of operations
        self.commit_index = 0

    def apply(self, operation):
        """Apply operation to state machine"""
        op_type = operation['type']

        if op_type == 'set':
            self.state[operation['key']] = operation['value']
        elif op_type == 'delete':
            self.state.pop(operation['key'], None)
        elif op_type == 'increment':
            key = operation['key']
            self.state[key] = self.state.get(key, 0) + 1

    def append_to_log(self, operation):
        """Add operation to log (not yet applied)"""
        self.log.append(operation)
        return len(self.log) - 1  # Return log index

    def commit(self, index):
        """Apply all operations up to index"""
        while self.commit_index <= index:
            self.apply(self.log[self.commit_index])
            self.commit_index += 1

    # The consensus protocol ensures all replicas
    # have the same log in the same order

Mental Model

Lamport approaches distributed systems by asking:

  1. What is the state? Define the state machine precisely
  2. What are the safety properties? What must never happen?
  3. What are the liveness properties? What must eventually happen?
  4. How do we order events? Logical time, not physical time
  5. Can I specify this formally? If not, I don't understand it

Signature Lamport Moves

  • TLA+ specifications before implementation
  • Logical clocks for event ordering
  • Safety and liveness as separate concerns
  • State machine replication for consensus
  • Happens-before reasoning
  • Formal proofs of correctness

平台分布

Codex

36%
按下载量换算29

Claude

32.36%
按下载量换算26

Cursor

19.33%
按下载量换算15

Gemini CLI

8.46%
按下载量换算7

安全审计

Gen Agent Trust Hub

未通过

Socket

通过

Snyk

通过

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。来源安全扫描存在提醒或未通过项,安装前请自行复核。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills