Skip to content

NocodeMrLi/mini-program-engineering-skill-suite / foundation

evidence-first-engineering

Verified

Apply evidence-first engineering discipline to any agent-built software project: fact discovery before action, evidence-calibrated status reporting, explicit change boundaries and stage gates, controlled confirmation for risky or external actions, sensitive-information redaction for anything shared publicly, and resumable continuity after interruptions. Use when users ask an agent to take over an existing codebase, deliver a feature across stages, verify whether work is actually complete, judge release readiness, or keep conclusions honest about what is proven versus assumed. This foundation skill is domain-neutral; vertical suites (for example mini-program engineering) build on it by adding domain facts and platform rules.

guardrails/engineering-guardrails.md

View on GitHub
# 共享工程门禁

所有组件在执行前读取并遵守本文件。门禁用于限制阶段跳跃、越界修改和无证据状态升级。

## 事实门禁

- 已有项目先读取规则文件、代码结构、相关文档、配置和当前版本控制状态。
- 当前代码、配置和平台返回优先于历史方案;截图只证明某一时刻的可见现象。
- 可能变化的平台规则、接口、价格或审核条件需要当前官方证据。
- 无法确认的内容标记为“假设”或“未知项”。

## 范围门禁

- 只修改与当前目标直接相关的文件、模块与行为,保留用户已有改动。
- 不为了界面完整度发明功能、入口、状态或数据规则。
- 发现相邻问题可以报告,但不能静默扩大实现范围。
- 用户要求删除某项时检查模板、样式、状态、计算和资源引用,避免只隐藏表象。

## 实现与调试门禁

- 沿用项目现有框架、工具、组件和数据模式,除非有证据支持变更。
- 先确认事实源,再清理重复资产或状态。
- 调试先复现;从日志、数据、默认值、异步时序、缓存和构建版本寻找根因。
- 修复覆盖同类状态与受影响契约,不用更长等待或隐藏错误掩盖失败。

## 验证与发布门禁

- 验证按风险覆盖静态、单元、集成、状态矩阵、真机、云端和发布层级。
- 只报告实际执行到的层级,自动化结果不等于用户正式验收。
- 发布前核对目标分支、最新构建、测试开关、调试入口、隐私与敏感信息。
- 没有平台侧证据时,不报告“已上传”或“已发布”。

## 恢复、资产与证据门禁

- 工具异常、超时、用户中止、会话断开或外部响应不明确后,在途动作先标记为 `unknown`;刷新工作区、进程、产物、日志和外部状态后再决定是否继续,禁止盲目重放。
- 处理图片、动画、音频、字体等资产时记录原始/衍生关系、处理方式、目标槽位、尺寸、透明通道、哈希、批准范围和替换关系;肉眼相似不等于同一资产。
- 证据进入结论前记录产生工具、格式、版本、时间、环境/版本指纹、适用结论和不能证明的内容;来源、版本或完整性不足时只作为受限线索。

## 来源项目独立性门禁

- 从真实来源项目蒸馏 Skill 时,来源项目只提供历史经验,不得作为该 Skill 的开发回归、前向验证、正式验收、整改或示范目标。
- 用来源项目验证其衍生 Skill 会形成循环验证,不能提供独立的有效性证据;历史来源项目结果不得计入冻结或发布门禁。
- Skill 验证必须使用匿名夹具、合成项目、隔离副本或与来源无关的授权项目,并确保写入始终限制在这些独立材料内。
- 不因 Skill 测试、验收或维护读取、构建、修改来源项目,也不在其中生成报告、缓存、产物或其他文件。
<!-- foundation-source: evidence-first-engineering v3.0 -->
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 471kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 322kUpdated Claude CodeCodexVerifiedInstall