Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/mileson/openprd/openprd-qualitynpx skills add mileson/openprd --skill openprd-qualitygit clone --depth 1 https://github.com/mileson/openprdWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/mileson/openprd/openprd-quality)<a href="https://agentmods.dev/skills/mileson/openprd/openprd-quality"><img src="https://agentmods.dev/badge/skills/mileson/openprd/openprd-quality.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00055 | $0.03349 |
| Opus 5 | $0.00028 | $0.01674 |
| Sonnet 5 | $0.00011 | $0.00670 |
| Haiku 4.5 | $0.00006 | $0.00335 |
Grade A, and why
openprd-quality scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 115 lines — stays where its author put it; the contents beside it link to each section on GitHub.
OpenPrd Quality
何时使用
当实现就绪、日志、链路追踪、业务成本、免费额度、滥用防护、评估执行环境、媒体证据、冒烟测试、性能阈值、压力数据、HTML 质量评估报告或项目级经验 Skill 在范围内时，使用这份 skill。
典型触发词：
- 日志、链路追踪、中心化日志、排查证据、报错回溯
- 免费用户、额度、限流、AI 调用、第三方 API、成本、预算、报警、止损、滥用
- eval、评估体系、冒烟测试、功能覆盖、异常流程、逆向流程
- 音频证据、media-flow、ASR/转写证据包、最终媒体复测
- CPU、内存、加载时间、接口耗时、压力测试、极端数据
- 质量评估报告、HTML 审查产物、质量门禁
- 界面效果图、实现截图、视觉对比、复刻对标、大界面改动方案评审、阶段性视觉评审
- 复盘后沉淀经验 Skill，避免同类问题反复出现
核心命令
- 初始化质量状态：
openprd quality <path> --init
- 生成评审产物：
openprd quality <path> --verify
- 先生成带来源的后台经验候选：
openprd quality <path> --learn --review --from .openprd/harness/turn-state.json
- 生成界面视觉对比图：
openprd visual-prepare <path> --reference <效果图> --grid <列>x<行>openprd visual-prepare <path> --reference <效果图> --boxes <plan.json>openprd visual-compare <path> --reference <效果图> --actual <实现截图> --locale <zh-CN|en>openprd visual-compare <path> --before <修改前截图> --after <修改后截图> --locale <zh-CN|en>openprd visual-compare <path> --board <focus-board.json|parallel-board.json|verification-board.json|alignment-board.json> --locale <zh-CN|en>
- 大界面改动的实现前方案评审：
- 先按用户目标、信息架构变化、视觉决策成本和验证风险判断方案评审形态，用户无需另行提出生图；已有界面时按平台能力截取当前真实界面，三个方向共享同一截图并默认保持原视觉 DNA；冷启动用已确认 PRD、用户群体、第一版切片、视觉目标、气质端点和记忆点生成 design brief；再按工具面用 Codex
imagegen（Image 2）或 CursorGenerateImage生成至少 3 个设计方向，并保存横向拼接评审大图到.openprd/harness/visual-reviews/
- 先按用户目标、信息架构变化、视觉决策成本和验证风险判断方案评审形态，用户无需另行提出生图；已有界面时按平台能力截取当前真实界面，三个方向共享同一截图并默认保持原视觉 DNA；冷启动用已确认 PRD、用户群体、第一版切片、视觉目标、气质端点和记忆点生成 design brief；再按工具面用 Codex
- 基于已审查报告生成或刷新项目级经验：
openprd quality <path> --learn --from <candidate-dir|report-id-or-json>
- 审查执行中发现的配置、规则候选或 user-local 偏好：
openprd grow <path> --review
质量审查契约
每次阶段性实现都应先判断 EVO 场景，再按场景要求审查。基础场景至少要覆盖冒烟和任务/功能覆盖；涉及 UI、桌面端、后端、成本、安全、性能、极端数据、发布交付时，再叠加对应门禁。
脚本、依赖或 fixture “存在”只能说明项目具备能力，不能说明本次已经执行。质量报告必须区分：
pass：当前场景必需门禁已经满足，且有本次执行证据或明确项目证据needs-evidence：能力可能存在，但缺本次执行证据needs-attention：能力、配置、任务或护栏本身缺失advisory：当前场景不阻断，但发布或风险进入范围时必须补齐
每次阶段性实现都应从六层审查：
- 质量契约：这项任务承诺覆盖什么
- 可观测性：前端、后端、agent 工具、异步任务和错误路径能否通过共享 trace/request/task/error id 串起来
- 业务成本与滥用护栏：免费、试用、消耗型资源、AI 调用、第三方 API、下载、存储等路径是否有额度、负向验证、监控、报警和止损
- 评估执行环境：冒烟测试、功能覆盖、正常性能和极端数据场景是否存在并持续维护
- 媒体证据：
media-flow任务是否通过现有test-strategy门聚合openprd.media-evidence.v1，并完成真实来源锁定、修改前基线、最终媒体重新提取与同口径复测；HTML、波形图或中间字幕只能承载审查，没有人工 gold transcript 时转写准确率保持 unknown - 视觉评审证据：大界面改动应存在 3 方向效果图横向评审大图，并记录 Agent 采用的可逆默认方向或用户明确选择及其来源；每个方向应有明确用途、受众、气质端点、审美主张和记忆点，并完成 anti-slop 自检。涉及界面视觉实现且已有用户参考效果图或 Agent 默认参考方向时，确认
.openprd/harness/visual-reviews/下存在本次openprd visual-compare输出的“效果图 / 实现截图”JPG，并且 Agent 已基于合成图复核差异；如果参考图是一张整板、网格图或多对象候选图，进一步确认已存在openprd visual-prepare产出的 reference-set、contact sheet 与 compare-plan 或 board 模板，并且 contact sheet 已被检查。没有参考图时先区分新建界面和修改既有界面，新建界面确认 Agent 已在后台完成 3 方向方案评审并采用可逆默认方向，修改既有界面确认存在“修改前 / 修改后”JPG，并已检查预期变化和未改区域漂移；当验收关注局部细节时，确认存在“局部焦点证据板”；当并行跑了多个优化方向时，确认存在“并行实验证据板”；当普通截图、Computer/Browser/Playwright 实测截图作为视觉证据时，确认存在“截图实测证据板”；当新功能或改动包含同构列表、卡片、网格或表格，或用户反馈没有对齐/排版漂移时，确认存在“对齐辅助线证据板”，并检查标题、标签、描述、状态、操作区等相同槽位的 x/y/宽高 spread 是否低于阈值。这些证据不只查有无图片，还要查气质、层级、字体/色彩/动效/表面角色和记忆点是否成立。 - 局部验收证据：如果任务只改一个组件或不超过 3 个相邻区域，局部裁剪图必须是主证据，不能只生成全屏对比后口头说已复核。证据 artifact 应记录
presentation=local-first、左右独立裁剪坐标、sharedScale、stretched=false、差异图和底部上下文；如果左右局部图被分别拉伸成同宽、裁剪的不是同一语义对象，或 Agent 没有根据差异重新截图返工，不能宣称局部视觉验收完成。 - 轻量 UI 可视优化：卡片宽度、间距、留白、对齐、颜色、圆角、字号、按钮或图标等小改仍属于用户可见变化。质量审查时至少确认存在修改前后视觉对比、局部焦点证据板、截图实测证据板或对齐辅助线证据板，并且本轮审美意图、气质、层级和记忆点已被复核；同构列表、卡片、网格或表格即使用户没有主动提出“对齐”，也要把相同文案类型/相同组件槽位的对齐作为默认验收项；build、package、
dev-check、单元测试或单张原始截图都不能替代视觉收口证据。 - HTML 质量评估报告：
.openprd/quality/reports/下的人类审查产物是否存在，且足以支持就绪判断；报告目录按reports.retentionMaxReports（默认 60）自动清理旧 eval 报告，index 保留的最新条目不受影响，不要依赖已被清理的历史报告路径 - 项目经验沉淀：已验证修复若具备复用价值，先生成带来源的当前项目经验候选并在后台记录“本次情况 / 计划保留的经验 / 以后怎么复用 / 只保留在当前项目里”；不要因此创建用户确认停顿
- 自我成长：配置缺口、文件识别、命令习惯或用户偏好优先沉淀为带来源的
.openprd/growth候选；高置信低风险项可后台固化，其他项保留候选，不要把个人偏好混进项目共享质量经验
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- yesterday First seen · 115 lines · 55 tokens per session scan A c492b9c779b1
openprd-quality is a skill published in the GitHub repository mileson/openprd (49 stars, last pushed 6d ago), licensed MIT. It adds 55 tokens to every session and 3,349 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
Other skills, from other repositories
openspec-bulk-archive-change
Archive multiple completed changes at once. Use when archiving several parallel changes.
openspec-explore
Enter explore mode - a thinking partner for exploring ideas, investigating problems, and clarifying requirements. Use when the user wants to think through something before or during a change.
openspec-onboard
Guided onboarding for OpenSpec - walk through a complete workflow cycle with narration and real codebase work.
release-openspec
Use this skill when releasing OpenSpec: audit merged work and changeset coverage, decide whether a catch-up changeset PR is needed, prepare or resume the Changesets Version Packages PR, cut a beta or stable release, verify publishing, and polish GitHub release notes. Also use when asked whether an open release PR is…
openspec-sync-specs
Sync delta specs from a change to main specs. Use when the user wants to update main specs with changes from a delta spec, without archiving the change.
openspec-archive-change
Archive a completed change in the experimental workflow. Use when the user wants to finalize and archive a change after implementation is complete.