Agent skill

Skill Lint

by cat-xierluo in cat-xierluo/legal-skills

Skill 创建预检、可靠性验收与格式审查工具。本技能应在用户创建、重大改造或审查Skill,需要识别旧版 Skill 的指令遵循不稳定、产出漂移、验证模态错配、约束漏检,或检查 Harness 契约、候选绑定证据、故障注入、目录结构、业务流和安全风险时使用。不要用于:代替业务领域验证器、代码审查、应用功能测试、通用编程任务。

MITAuto-check: notesDevelopment

Install Skill Lint

skills CLI
$ npx skills add cat-xierluo/legal-skills --skill skill-lint -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install cat-xierluo/legal-skills skill-lint --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/cat-xierluo/legal-skills.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/skill-lint .claude/skills/skill-lint && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
skill-lint
GitHub stars
720
Token cost
~2.9k tokens
SKILL.md length
553 words
Files
36 (incl. scripts, references)
Skills in repo
62
Repo updated
First seen
Licence
MIT

At a glance

Skill 创建预检、可靠性验收与格式审查工具。本技能应在用户创建、重大改造或审查Skill,需要识别旧版 Skill 的指令遵循不稳定、产出漂移、验证模态错配、约束漏检,或检查 Harness 契约、候选绑定证据、故障注入、目录结构、业务流和安全风险时使用。不要用于:代替业务领域验证器、代码审查、应用功能测试、通用编程任务。

  • Works in 9 steps: 选择模式 → 确认范围 → 扫描文件 → …
  • Tasks that involve Linting and formatting
  • SKILL.md covers 工作原则, 输入, 审查流程 and 问题分级, plus 2 more sections
  • Calls python3

What it does

Skill Lint is an agent skill from cat-xierluo/legal-skills. Skill 创建预检、可靠性验收与格式审查工具。本技能应在用户创建、重大改造或审查Skill,需要识别旧版 Skill 的指令遵循不稳定、产出漂移、验证模态错配、约束漏检,或检查 Harness 契约、候选绑定证据、故障注入、目录结构、业务流和安全风险时使用。不要用于:代替业务领域验证器、代码审查、应用功能测试、通用编程任务。

Its SKILL.md is about 2.9k tokens, which your agent loads only when the skill is triggered. The skill folder holds 38 other files, including scripts and reference files (for example `CHANGELOG.md`, `config/instruction-stability-contract.example.json` and `config/instruction-stability-held-out-cases.example.json`).

It sits in Development, covering Linting and formatting. The licence is MIT.

When your agent uses it

  • Tasks that involve Linting and formatting

Example prompts

  • “/skill-lint”

Requirements

  • Python 3

Workflow steps

9 steps, taken from the step headings in SKILL.md.

  1. 选择模式
  2. 确认范围
  3. 扫描文件
  4. 模块化规则审查
  5. 安全性评估
  6. 业务流深度审查
  7. Harness 可靠性审查
  8. 可评估性审查
  9. 生成审查报告

What it can do on your machine

Read from SKILL.md and the folder at commit db2c58c. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Ships 1 file in scripts/, which the agent can run.

    Shell commands in SKILL.md call:

    • python3

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Skill Lint loads about 2.9k tokens when it runs, and up to ~39k if it reads all its reference files. Until then it costs about 44 tokens; SKILL.md has 553 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~44
When it runs · the whole SKILL.md, loaded when a task matches
~2.9k
With references · SKILL.md plus every file in references/, read only if the agent opens them
~39k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check: notes

The automated check noted patterns worth knowing about, such as sudo or a known installer.

  • NoteMentions a .env fileSKILL.md:89
    如果在 Skill 单元内出现 `.env`、真实密钥、`__pycache__/`、`docs/`、`test/` 等发布版不应包含的内容,按严重程度记录。仓库根目录的 README、docs、LICENSE、CHANGELOG 可以是

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); the scripts in this folder are not scanned.

SKILL.md

The full file from cat-xierluo/legal-skills at commit db2c58c, republished under its MIT licence (© cat-xierluo). 553 words, ~2,930 tokens.

Download SKILL.mdSave it as .claude/skills/skill-lint/SKILL.md (or your agent's skills folder). This skill also uses 35 other files; get the full folder from GitHub.
name
skill-lint
description
Skill 创建预检、可靠性验收与格式审查工具。本技能应在用户创建、重大改造或审查Skill,需要识别旧版 Skill 的指令遵循不稳定、产出漂移、验证模态错配、约束漏检,或检查 Harness 契约、候选绑定证据、故障注入、目录结构、业务流和安全风险时使用。不要用于:代替业务领域验证器、代码审查、应用功能测试、通用编程任务。
homepage
https://github.com/cat-xierluo/legal-skills
author
杨卫薪律师(微信ywxlaw)
version
2.9.1
license
MIT

Skill Lint

本技能负责 Skill 创建前设计预检和创建后质量验收:审查一个 Skill 是否结构合规、文档一致、可发布、可评估、安全风险可控,判断它是否真实承载业务流程,并检查“完成”结论是否由候选绑定证据和多轮真实产物支撑。

本技能不代替主要创建者或领域验证器。创建或大改 Skill 时,用它先审查 Harness 设计,再由创建工具实现,完成后回到本技能做候选绑定验收。

工作原则

  • 先看硬性问题,再看优化问题。
  • 先做静态审查,再判断业务流深度。
  • 先定位审查单元,再审查目标 Skill 目录及明确给出的上下文。
  • 不把格式合规等同于任务效果通过。
  • 不采信生产者自报的 PASS;正式验收必须绑定当前候选、当前规则和实际日志。
  • 不把“有 checker”当作“覆盖完整”;每条硬约束必须追踪到合适模态、正确产物阶段和回归用例。
  • 不比较自然语言输出的整文件哈希;只比较合同声明的关键覆盖集合和可观察不变量。
  • 客观缺陷 fail-closed;语义质量保留人工判断,不伪装成万能自动化。
  • 只审查领域 checker 的覆盖、模态和证据,不把证据不足冒充为已经发现具体领域算法错误。
  • 已知漏报必须有最小违规反例,已知误报必须有合法近似正例;两者都由目标 Skill 的领域 checker 定义。
  • 对无法确认的能力标注“未提及/待补充”。

输入

审查时至少需要:

  • 目标路径或仓库地址:可以是单个 Skill 目录、monorepo 根目录、GitHub 仓库或待改造的提示词集合
  • 审查目的:创建前设计预检、发布前验收、改造评估、他人 Skill 审查、回归检查等

可选输入:

  • 用户给出的特殊偏好或项目规则
  • 本地审查配置文件,如 config/review-profile.local.yaml
  • 需要重点关注的问题清单

如需配置个人或项目的发布元数据策略,先复制 config/review-profile.example.yaml 为 config/review-profile.local.yaml,再填入本地值。个人偏好只作为本地上下文使用,不写入公开文件,不复制到审查报告中,除非用户明确要求公开。

审查流程

0. 选择模式
  • 创建预检:目标尚未实现或将重大改造。先读取 references/harness-reliability-standards.md,产出七层 Harness 设计、Hard Fail 和至少一个逃逸反例;不要直接扩大提示词。
  • 快速审查:检查第三方 Skill、草稿或局部问题。可以只做静态与语义审查,但结论必须写 NOT_VERIFIED,不得称功能已验收。
  • 旧版稳定性审查:检查老版本、多维 review、视觉生产或反复漏项的 Skill。先读取 references/instruction-stability-standards.md 并运行静态 assess;缺少约束追踪合同或多轮证据时标记 NOT_VERIFIED,不得用一次成功执行推断稳定。
  • 正式验收:发布、交付或声称“稳定完成”前,除全部审查模块外,还必须运行候选绑定证据门禁。动态执行只用于用户已确认的自有/可信候选;未知第三方候选默认停在 NOT_VERIFIED,除非用户明确授权并使用隔离环境。

创建预检完成后,由用户指定的创建工具或实现者落地;本技能在实现完成后重新进入正式验收。这样既把可靠性理念前移,又不让审查器与生产器混成同一责任。

1. 确认范围

先读取 references/repository-skill-discovery-standards.md,判断输入目标是哪一类:

  • 单个 Skill 目录:目标目录自身包含 SKILL.md
  • monorepo / Skill 集合:仓库根目录只是容器,内部多个子目录才是最小 Skill 单元
  • 松散提示词集合:没有标准 SKILL.md 单元,但存在带 name / description frontmatter 的 Markdown 或 README 索引
  • 普通仓库:没有足够证据表明包含 Skill

不要只因为仓库根目录缺少 SKILL.md 就判定整个仓库不合格。根目录缺少 SKILL.md 只有在用户明确指定根目录就是单个 Skill,或仓库声明自己是一个可加载 Skill 根目录时,才按严重问题处理。

发现候选单元后,先列出:

  • 已确认 Skill 单元:目录内有 SKILL.md
  • 非标准但可迁移的 Skill-like 文档:单个 Markdown 带 name / description frontmatter,或 README 明确称为 skill
  • 仓库级治理文件:README、LICENSE、CHANGELOG、Marketplace、贡献说明

如果候选单元很多,先按用户指定范围审查;用户未指定时,优先审查已确认 Skill 单元,并抽样检查 Skill-like 文档,报告中说明抽样范围。

2. 扫描文件

对每个已确认或被选中的候选单元列出文件,并重点检查:

  • SKILL.md
  • CHANGELOG.md
  • LICENSE.txt
  • config/*.example.*
  • references/*.md
  • scripts/*
  • assets/*
  • templates/*
  • archive/.gitkeep

如果在 Skill 单元内出现 .env、真实密钥、__pycache__/、docs/、test/ 等发布版不应包含的内容,按严重程度记录。仓库根目录的 README、docs、LICENSE、CHANGELOG 可以是 monorepo 治理文件,不按单个 Skill 目录结构误判。

3. 模块化规则审查

先读取 references/skill-standards.md 作为审查索引,再按问题类型读取对应模块。不要一次性把所有细则混在一份报告逻辑中。

默认模块:

  • repository-skill-discovery-standards.md:仓库类型、monorepo、最小 Skill 单元和候选文档发现
  • structure-standards.md:目录结构、文件可达性、references 命名
  • frontmatter-metadata-policy.md:通用字段与发布字段分层
  • trigger-description-standards.md:name 与 description 触发边界
  • configuration-privacy-standards.md:配置模板、本地配置隔离、公开内容去具体化
  • security-assessment-standards.md:危险执行、敏感访问、数据外传、凭证、依赖、MCP 和提示词安全
  • publishing-standards.md:LICENSE、CHANGELOG、version、README / marketplace 同步
  • workflow-output-standards.md:SKILL.md 正文、依赖、脚本、输出和可编排性
  • business-flow-rubric.md:业务流深度、Hard Fail 和可评估性基础
  • harness-reliability-standards.md:七层 Harness、独立验证、候选绑定证据、故障注入和闭环
  • instruction-stability-standards.md:约束追踪、验证模态、产物阶段、多轮覆盖和漂移判定
  • reporting-standards.md:问题分级和报告结构

LICENSE.txt、version、README 和 Marketplace 属于发布治理,不属于普通目录结构硬要求。审查私人或第三方普通 Skill 时,只有在用户给出发布目标或项目规则时才按发布模块判定。

4. 安全性评估

读取 references/security-assessment-standards.md,对纳入审查的 Skill 单元做安全风险评估。

先运行确定性静态安全扫描(对应 NVIDIA SkillSpector 的漏洞模式):

bash
# 单个 Skill
python3 scripts/security_scan.py audit --candidate-root /path/to/skill

# Skill 集合 / monorepo(递归发现所有含 SKILL.md 的最小单元)
python3 scripts/security_scan.py batch --root /path/to/skills

# 联网查询 OSV 已知漏洞(默认离线只做版本 pin 检查)
python3 scripts/security_scan.py audit --candidate-root /path/to/skill --online

扫描器输出含文件、行号、能力信号、严重级、置信度与修正建议的结构化 JSON;存在 critical/high finding 时退出码 1(FAIL),未发现 SKILL.md 时退出码 2。覆盖:危险执行(subprocess/os.system/eval/exec/动态导入)、网络外传、环境变量/敏感文件访问、运行时自动安装、未固定依赖与已知 CVE、硬编码凭证、提示注入、污点流(env/argv/input → subprocess)、文件系统枚举、隐藏 Unicode 字符、MCP 通配权限,以及"能力存在但文档未披露/未声明权限"(Missing User Warnings / MCP Least Privilege / Context-Inappropriate Capability)与文档级 scope creep(commit 技能引导发布/改 allowlist 等未披露的高风险动作)。

重点检查:

  • SKILL.md 和 references 是否含提示注入、绕过安全限制、隐藏执行、敏感数据收集或欺骗性描述
  • scripts 是否含危险命令执行、下载并执行、权限提升、无边界删除、敏感文件访问、数据外传、动态导入或混淆
  • config/example 是否含真实凭证、真实 endpoint、真实 webhook 或本地敏感路径
  • 依赖、安装钩子、MCP、网络请求和外部工具权限是否有用途说明、范围限制和用户确认
  • GitHub 仓库审查时,提交历史是否出现过敏感信息泄露、异常删除重加或与 Skill 行为不一致的提交

安全评估不等同于完整渗透测试。对扫描器命中项要结合上下文判断误报;但涉及凭证泄露、下载并执行、权限提升、持久化、无确认数据外传、隐藏提示词指令等问题时,默认按严重问题处理。

5. 业务流深度审查

使用 references/business-flow-rubric.md 检查:

  • Trigger:是否清楚说明何时触发、何时不触发
  • Intake:是否识别输入缺口并规定追问方式
  • Reasoning:是否区分事实、归纳、判断和依据
  • Output:是否定义输出结构、验收标准和后续动作
  • Safety:是否控制隐私、过度承诺和高风险场景

默认采用中等严格度:Hard Fail 是硬指标,五层评估对象是软指标。

Show full SKILL.md (296 more words)Show less
6. Harness 可靠性审查

创建预检、重大改造和正式验收必须读取 references/harness-reliability-standards.md,逐层检查 Contract / Producer / Verifier / Evidence Binding / Fault Injection / Closure / Composition。旧版、多维审阅、视觉生产或用户反馈“反复漏项/产出漂移”时还必须读取 references/instruction-stability-standards.md。

先运行不执行候选代码的具体失效审查。单 Skill:

bash
python3 scripts/harness_failure_audit.py audit \
  --candidate-root /path/to/skill

Skill 集合或 monorepo:

bash
python3 scripts/harness_failure_audit.py batch \
  --root /path/to/skills

该审查器递归发现包含 SKILL.md 的最小单元,输出带文件、行号、证据、影响和修正建议的稳定 JSON;hard finding 返回 1,范围错误或集合中零 Skill 返回 2。规则 ID 与边界见 references/harness-reliability-standards.md。默认只做静态审查,不执行、安装或联网。

旧版结构性稳定性审查:

bash
python3 scripts/instruction_stability_gate.py assess \
  --candidate-root /path/to/skill

缺少 config/instruction-stability-contract.json 时,assess 返回退出码 2 和 INSTRUCTION_STABILITY_NOT_VERIFIED,同时给出 ISG 结构性 finding 与 HFA/HRA 具体实现 finding。视觉/几何语义只从 SKILL.md 和 references/**/*.md 的规范性上下文识别,TASKS/DECISIONS/CHANGELOG 的历史讨论不触发视觉模态。此模式不执行候选代码,适合旧版和未知第三方 Skill。

ISG-002 会列出触发视觉/几何要求的规范来源和行号,但它只证明候选缺少合适模态的领域验证证据。不要据此声称已经检测出重叠、裁切或其他具体业务错误;应让目标 Skill 提供自己的 checker、最小违规反例与合法近似正例,再由正式门禁复算。

声称“指令遵循稳定”“多轮不漏项”前,目标 Skill 必须提供约束追踪合同。每条 hard constraint 在权威来源中使用唯一 <!-- skill-lint:constraint CONSTRAINT-ID --> 锚点;候选外 evaluator-signed 基线必须与全部锚点、规范行、合同和当前候选哈希一致。先取得当前候选可复算的 Harness 审查证据,再用相同输入/配置至少独立执行三轮;每轮保留唯一 execution nonce、evaluator-signed producer log 和独立目录内的真实产物。每条硬约束还必须有候选外 evaluator-signed held-out 正反例。完成静态安全审查、披露 checker 且用户确认候选为自有/可信代码后,运行:

bash
python3 scripts/instruction_stability_gate.py verify \
  --candidate-root /path/to/skill \
  --evaluator-public-key /path/to/review/evaluator-public.pem \
  --requirements-baseline /path/to/review/requirements-baseline.json \
  --harness-evidence /path/to/review/harness-review.json \
  --held-out-cases /path/to/review/held-out-cases.json \
  --held-out-root /path/to/review/held-out \
  --run-evidence /path/to/review/runs.json \
  --runs-root /path/to/review/runs \
  --receipt /path/to/review/instruction-stability-receipt-draft.json \
  --confirm-trusted-candidate

该门禁固定使用当前受信 skill-lint 复算 HARNESS_REVIEW_VERIFIED,再逐轮重跑 active checker。它要求每条硬约束精确映射到显式来源锚点、自动发现并签名枚举的 requirements sources/exclusions、checker、正确验证模态、正确产物阶段、带类型/条件/阈值的 measurement、签名 hidden 正反例和已知历史回归;合同漏列规范行、签名无效、三轮复用路径、旧候选/旧 producer 日志重放、负向用例因其他约束失败、少报一个 constraint id、checker 修改产物或关键 observable 漂移都会阻塞。公开、隐藏和真实运行产物统一使用同类随机暂存路径,避免 checker 按样本类别分支。

verify 只生成 INSTRUCTION_STABILITY_EVIDENCE_READY 草稿。evaluator 必须在不执行候选代码的隔离环境用 Ed25519 私钥运行 sign-evidence --private-key ...,再用只持有候选外公钥的 verify-receipt 复验全部绑定;只有后一步输出 INSTRUCTION_STABILITY_VERIFIED 才能声明稳定。私钥不得出现在 producer/checker 的进程树、环境或可读工作区,完整命令见 references/instruction-stability-standards.md。

客观 Hard Fail 一律阻塞。正式验收应先用 scripts/harness_evidence_gate.py snapshot 固化候选与规则读集,填写候选内 checker、参数、超时和故障用例的预期失败码。完成静态安全审查并取得用户对自有/可信候选的确认后,再用 verify --confirm-trusted-candidate 亲自重跑。JSON 中不得填写或采信自报退出码、PASS 和日志。只有退出码为 0 且出现 HARNESS_REVIEW_VERIFIED,才能说“当前候选的 Harness 审查证据已验证”。该标记不替代目标 Skill 自己的 DOMAIN_VERIFIED,门禁也不是第三方代码沙箱。

完成标记不可互相替代:

  • HARNESS_REVIEW_VERIFIED:当前候选、规则、checker 和故障用例证据有效。
  • INSTRUCTION_STABILITY_EVIDENCE_READY:动态复算已完成,但草稿尚未离线签名与验签,不能作为完成标记。
  • INSTRUCTION_STABILITY_VERIFIED:evaluator Ed25519 签名回执已复验,外部硬约束基线/held-out、当前 Harness evidence、候选/producer 绑定和至少三轮真实产物逐约束通过,measurement/observable 未漂移。
  • DOMAIN_VERIFIED:领域验证器确认具体业务产物正确。

只要 Skill 声称“稳定完成”,就必须同时满足前两项;若还声称业务结果正确,再加第三项。未运行的层一律写 NOT_VERIFIED。

7. 可评估性审查

确认 Skill 是否具备后续 eval 的基础:

  • 是否声明评估范围
  • 是否声明 Hard Fail
  • 是否提供 benchmark case 或样例
  • 是否提供输出验收标准
  • 是否区分静态检查与动态评估
  • 是否有逐约束追踪合同、验证模态和产物阶段
  • 是否用至少三轮固定输入检查关键覆盖集合与历史回归
  • 已知漏报是否有最小违规反例,已知误报是否有合法近似正例

缺少这些内容不一定阻塞发布,但应作为质量风险记录。

8. 生成审查报告

审查报告应优先列出问题,再给摘要。严重问题必须具体到文件和位置。

如用户需要最终交付件、发布前意见或正式质量结论,使用 templates/skill-quality-opinion-report.md 生成“Skill 质量意见报告”,报告中必须写明问题、影响、修正方式和复查标准。

对承载设计原理的结构性建议(拆解披露、触发边界、上下文聚焦、自由度匹配、可机判验收等),在 finding 的「设计理念」字段一句话讲清背后写作原理,可回查对应 standards 文件的「设计理念」小节,使报告同时具备 skill 写作教学价值;纯事实问题(文件缺失、引用断裂、命名大小写)可省。稳定性 finding 必须具体指出漏的是哪条 constraint、验证模态、产物阶段、case 或 run,不使用“遵循不稳定”这类无法复查的泛称。

生成正式质量意见报告后,按 references/archive-standards.md 判断是否归档。需要归档时,在本技能 archive/YYYYMMDD_HHMMSS_<target-slug>/ 下保存报告、元数据和证据索引;真实归档内容不提交到 Git。

问题分级

级别说明处理
❌ 严重阻塞加载、发布、使用安全或质量验收必须修复
⚠️ 警告影响维护、复用、审查可信度或可评估性建议修复
ℹ️ 信息风格、清晰度或后续改进建议可选处理

Hard Fail 一律按严重问题处理。

报告模板

正式质量意见报告模板见 templates/skill-quality-opinion-report.md。报告除原有结构、安全、业务流和可评估性外,必须单列 Harness 七层结论、证据等级和完成标记。

模板的设计理念要点(结构性建议必填「设计理念」字段,纯事实问题可省):

  • 严重问题:位置 + 依据 + 影响 + 修正方式 + 设计理念 + 最优设计 + 复查标准
  • 警告问题:位置 + 影响 + 建议修正 + 设计理念 + 最优设计 + 优先级
  • 安全评估:风险类别 + 安全级别 + 上下文(避免纯关键词误报)
  • 复查清单:严重问题已关闭 + 警告问题处理或登记 + 文件引用可达 + 隐私脱敏 + 安全覆盖 + 发布治理同步

参考规则

  • references/skill-standards.md:审查索引和模块路由
  • references/repository-skill-discovery-standards.md:仓库类型识别、monorepo 单元发现和候选文档分级
  • references/structure-standards.md:目录结构、文件可达性和 references 命名
  • references/frontmatter-metadata-policy.md:Frontmatter 通用字段与项目发布字段分层策略
  • references/trigger-description-standards.md:name 与 description 触发边界
  • references/configuration-privacy-standards.md:配置模板、本地配置隔离和公开内容去具体化
  • references/security-assessment-standards.md:危险执行、敏感访问、数据外传、凭证、依赖、MCP 和提示词安全
  • references/publishing-standards.md:LICENSE、CHANGELOG、version 与发布索引
  • references/workflow-output-standards.md:正文工作流、依赖、脚本、输出和可编排性
  • references/business-flow-rubric.md:业务流深度和可评估性判则
  • references/harness-reliability-standards.md:Harness 七层可靠性、Hard Fail、创建预检和候选绑定证据门禁
  • references/instruction-stability-standards.md:旧版识别、约束追踪、模态/阶段匹配、多轮执行和产出漂移门禁
  • references/reporting-standards.md:问题分级和审查报告模板
  • references/archive-standards.md:正式审查报告的内部归档机制
  • references/skill-dev-guide.md:Skill 开发规范参考
  • references/skill-orchestration-guide.md:复杂编排规范参考
  • config/review-profile.example.yaml:个人/项目审查配置模板
  • config/instruction-stability-contract.example.json:目标 Skill 的约束追踪合同示例
  • config/instruction-stability-requirements-baseline.example.json:候选外独立硬约束基线格式示例
  • config/instruction-stability-held-out-cases.example.json:evaluator-signed 候选外隐藏正反例清单格式
  • scripts/harness_evidence_gate.py:生成和复算候选绑定的 Harness 审查证据
  • scripts/harness_failure_audit.py:单 Skill / Skill 集合的具体 Harness 失效模式静态审查
  • scripts/instruction_stability_gate.py:静态识别旧版结构风险,并验证多轮真实产物的逐约束覆盖稳定性
  • templates/skill-quality-opinion-report.md:最终 Skill 质量意见报告模板

© cat-xierluo, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 35 other files (scripts, references) in skills/skill-lint of cat-xierluo/legal-skills.

  • SKILL.md
  • CHANGELOG.md
  • LICENSE.txt
  • archive/.gitkeep
  • config/instruction-stability-contract.example.json
  • config/instruction-stability-held-out-cases.example.json
  • config/instruction-stability-requirements-baseline.example.json
  • config/review-profile.example.yaml
  • references/archive-standards.md
  • references/business-flow-rubric.md
  • references/configuration-privacy-standards.md
  • references/frontmatter-metadata-policy.md
  • references/harness-reliability-standards.md
  • references/instruction-stability-standards.md
  • references/publishing-standards.md
  • references/reporting-standards.md
  • references/repository-skill-discovery-standards.md
  • references/security-assessment-standards.md
  • … and 18 more

Open the folder on GitHubat commit db2c58c

Compare with similar skills

Skill Lint next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Skill Lint compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Skill Lint this skillcat-xierluo/legal-skills720—~2.9kAutomated safety check: NotesMIT
Minimizing Ty Ecosystem Changesastral-sh/ruff50k—~4.6kAutomated safety check: PassMIT
Install Anti-Slop Oxlint Rulesdmmulroy/anti-slop5.4k—~2.2kAutomated safety check: PassMIT
Babysit PR To Pass CIsgl-project/sglang37k2 repos~3kAutomated safety check: PassApache-2.0
Rust Best Practicesfarm-fe/farm5.6k3 repos~1.1kAutomated safety check: PassMIT
Summarise Ecosystem Resultsastral-sh/ruff50k—~2.2kAutomated safety check: PassMIT

Similar skills

  • Official

    A skill your agent uses when a user says "minimize this ty ecosystem change", "reproduce this ecosystem result", "investigate a primer difference", "investigate a mypyprimer difference"…

    50k GitHub stars~4.6k tokensUpdated today
    DevelopmentAuto-check passed
  • Installs, updates or migrates the vendored anti-slop Oxlint plugin in a repository, keeping local rule changes and the plugin's license and provenance files.

    5.4k GitHub stars~2.2k tokensUpdated 1 mo ago
    DevelopmentAuto-check passed
  • Babysit PR To Pass CI

    sgl-project/sglang

    Start and persistently pursue a goal to babysit an SGLang pull request until selected GitHub Actions workflows pass on the latest PR head.

    37k GitHub starsUsed in 2 repos~3k tokens
    DevelopmentAuto-check passed
  • Guide for writing idiomatic Rust code based on Apollo GraphQL's best practices handbook.

    5.6k GitHub starsUsed in 3 repos~1.1k tokens
    DevelopmentAuto-check passed
  • Official

    A skill your agent uses when a user says "summarise ecosystem results", "summarize this ty ecosystem report", "what changed in this ecosystem run?", or asks to summarise or summarize ty ecosystem…

    50k GitHub stars~2.2k tokensUpdated today
    DevelopmentAuto-check passed
  • Go Pedantry

    chromedp/chromedp

    This skill should be used when the user is writing Go code and needs guidance on Go-specific pedantry: error wrapping with fmt.Errorf and %w, interface design (accept interfaces return structs)…

    13k GitHub stars~3.7k tokensUpdated 4 days ago
    DevelopmentAuto-check passed

More from cat-xierluo/legal-skills

All 62 skills in this repo
  • Elements-Style Complaint Generator

    cat-xierluo/legal-skills

    Converts a lawyer's ordinary complaint or a described case into the Supreme People's Court's elements-style Word template, with layout checks on the result.

    720 GitHub stars~2.5k tokensUpdated today
    Auto-check: notes
  • Lecture Performance Review

    cat-xierluo/legal-skills

    Analyzes raw lecture transcripts for verbal tics, pacing, time use and promise follow-through, with optional slide-by-slide comparison and cross-session tracking.

    720 GitHub stars~2.2k tokensUpdated today
    Auto-check passed
  • De-AI Polish for Chinese Articles

    cat-xierluo/legal-skills

    Detects and rewrites machine-sounding patterns in the body text of Chinese articles while keeping the author's facts, headings and legal terms intact.

    720 GitHub stars~2.9k tokensUpdated today
    Auto-check passed
  • GitHub Star Manager

    cat-xierluo/legal-skills

    Finds GitHub projects mentioned in articles or screenshots and stars them, tracks updates to your starred repos, and builds an HTML dashboard to browse them.

    720 GitHub stars~2k tokensUpdated today
    Auto-check: notes
  • Legal Harness Initializer

    cat-xierluo/legal-skills

    Sets up or incrementally updates AGENTS.md and CLAUDE.md for legal professionals, with a minimal safety baseline and a check that a new session loads and follows the rules.

    720 GitHub stars~2.3k tokensUpdated today
    Auto-check passed
  • Moot Court Simulation Builder

    cat-xierluo/legal-skills

    Chinese-language skill that organizes a case file into a multi-role mock trial with judge, parties and clerk, producing a transcript, issue review and a to-strengthen list.

    720 GitHub stars~1.4k tokensUpdated today
    Auto-check passed

Categories

Questions about Skill Lint

What does Skill Lint do?

Skill 创建预检、可靠性验收与格式审查工具。本技能应在用户创建、重大改造或审查Skill,需要识别旧版 Skill 的指令遵循不稳定、产出漂移、验证模态错配、约束漏检,或检查 Harness 契约、候选绑定证据、故障注入、目录结构、业务流和安全风险时使用。不要用于:代替业务领域验证器、代码审查、应用功能测试、通用编程任务。. Skill Lint is an agent skill from cat-xierluo/legal-skills.

When should I use Skill Lint?

Skill Lint fits situations like: tasks that involve Linting and formatting.

How do I install Skill Lint in Claude Code?

Run `npx skills add cat-xierluo/legal-skills --skill skill-lint -a claude-code`. Or copy the skill folder (skills/skill-lint in cat-xierluo/legal-skills) into .claude/skills/skill-lint in your project. Claude Code loads it when a task matches its description.

How do I install Skill Lint in Codex?

Run `npx skills add cat-xierluo/legal-skills --skill skill-lint -a codex`. Or copy the skill folder (skills/skill-lint in cat-xierluo/legal-skills) into .agents/skills/skill-lint in your project. Codex loads it when a task matches its description.

Can I use Skill Lint in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add cat-xierluo/legal-skills --skill skill-lint -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/skill-lint, .gemini/skills/skill-lint, .github/skills/skill-lint and .opencode/skills/skill-lint in your project.

What does Skill Lint need to run?

Going by SKILL.md and its folder, Skill Lint needs the command-line tools its instructions call (python3). Our summary lists: Python 3.

Does Skill Lint access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Skill Lint safe to install?

Our automated static check of SKILL.md found notes only (mentions a .env file), nothing it rates as a warning. It is not a guarantee. The check reads SKILL.md only: the scripts in the folder are not scanned, so read them before running anything.

What licence does Skill Lint use?

Skill Lint is published under the MIT licence (declared in SKILL.md). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Skill Lint use?

About 2.9k tokens (SKILL.md is roughly 12k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 36k tokens, read only when the agent opens those files.

What are the alternatives to Skill Lint?

Skills that share tags, products or a category with Skill Lint: Minimizing Ty Ecosystem Changes (astral-sh/ruff, 50k stars), Install Anti-Slop Oxlint Rules (dmmulroy/anti-slop, 5.4k stars), Babysit PR To Pass CI (sgl-project/sglang, 37k stars) and Rust Best Practices (farm-fe/farm, 5.6k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Skill Lint?

cat-xierluo (a GitHub user) maintains it in cat-xierluo/legal-skills, which has 720 GitHub stars. The repository holds 62 skills in this directory. The repository was last updated on October 10, 2026.

Source: cat-xierluo/legal-skills on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.