English

Execution Contract

Default to standard; explicit invocation is required. Complete one primary task by default. Continue an authorized end-to-end workflow within each Skill's permissions; Router remains recommendation-only. Pause for new material decisions or scope, not authorization already given. For a human-only blocker, give location, action, success signal and resume step; never request secret values or assume completion. Validate according to actual risk; security, production, migrations, releases and other high-risk work use audit. Read references/execution-modes.en.md only when mode selection is unclear.

Architecture Decision Record

Create or update one ADR when the user explicitly authorizes architecture-document edits. Do not modify business code.

Boundary

• Route requirements to requirements-to-spec, implementation sequencing to safe-change-plan, and technical risk reviews to the relevant specialized skill.
• Search for an existing ADR directory, index, numbering scheme, status vocabulary, template, and supersession rules before writing.
• Never present a proposal as accepted consensus. Use Proposed when approval evidence is absent.
• Use official standards only when allowed and necessary, recording source and access date; mark unavailable evidence unverified.

Workflow

1. Identify the real problem, scope, stakeholders, constraints, decision drivers, evidence, and review deadline.
2. Find related ADRs and determine whether to create, amend, supersede, or leave unchanged.
3. Compare viable alternatives, including status quo, with benefits, costs, risks, reversibility, compatibility, operational impact, and evidence.
4. Record status, date, context, drivers, options, decision, rationale, positive/negative consequences, risks, validation, and review conditions.
5. Preserve repository naming, location, numbering, links, language, and template. If none exists, use assets/adr-template.md.
6. Modify only the ADR or architecture index explicitly in scope. Report assumptions, unresolved questions, confidence, and implementation handoff without executing it.

Output Contract

1. Target ADR path and action
2. Evidence and related decisions
3. Status, date, context, and decision drivers
4. Options and tradeoffs
5. Decision and rationale, or Proposed recommendation
6. Consequences, risks, confidence, validation, and review conditions
7. Unknowns and implementation handoff

简体中文

执行契约

默认使用 standard；仅允许用户显式调用。 默认完成一个主任务。用户已授权完整流程时，在各 Skill 权限内连续完成；Router 仍只推荐。新重大决策或超出授权时暂停，不重复索取已有授权。 必须由用户完成的阻塞步骤：给出操作位置、具体动作、成功标志、恢复步骤；不要求粘贴秘密值，不代替用户确认完成。 按实际风险选择验证；安全、生产、迁移、发布等高风险工作使用 audit。模式不明确时才读取 references/execution-modes.zh-CN.md。

架构决策记录

仅在用户明确授权架构文档修改时创建或更新一个 ADR，不修改业务代码。

职责边界

• 需求交给 requirements-to-spec，实施顺序交给 safe-change-plan，技术风险交给相应专项 Skill。
• 写入前搜索现有 ADR 目录、索引、编号、状态词、模板和替代规则。
• 不得把提案写成已接受共识；没有批准证据时使用 Proposed。
• 仅在允许且必要时使用官方标准，记录来源和访问日期；不可访问时标为未验证。

工作流程

1. 识别真实问题、范围、利益相关者、约束、决策驱动、证据和复审时间。
2. 查找相关 ADR，判断创建、修订、替代或保持不变。
3. 比较可行候选及维持现状，覆盖收益、成本、风险、可逆性、兼容性、运维影响和证据。
4. 记录状态、日期、背景、驱动、方案、决策、理由、正负后果、风险、验证和复审条件。
5. 保留仓库命名、目录、编号、链接、语言和模板；不存在时使用 assets/adr-template.md。
6. 只修改范围内 ADR 或架构索引，报告假设、未决问题、置信度和实施转交，不执行实施。

输出契约

1. 目标 ADR 路径与动作
2. 证据与相关决策
3. 状态、日期、背景与决策驱动
4. 候选方案与权衡
5. 决策与理由，或 Proposed 建议
6. 后果、风险、置信度、验证与复审条件
7. 未知项与实施转交