@abelxiaoxing/cadence 1.4.0 → 1.4.1

package/README.md

@@ -26,6 +26,14 @@ Implement 只暴露 `start`、`status`、`resume`、`rebind`、`cancel` 和 `dis
 Design 从第一步起统一使用 `action: "design"`：新需求通过 `start(requirement)` 进入，已有 change 通过 `start(change)` 进入，后续只携带返回的 `runId`。
 需求、决策合同与 Gate A 合同由控制面规范化并计算哈希，调用方无需 SHA-256 工具；Gate B 自动绑定当前已编译 canonical plan，原始瞬时文本不会写入 durable journal。
 `validate-plan-draft` 可在 `compile-plan` 前只读运行同一套编译检查，并以结构化 `taskId` / phase / field / verification 诊断定位问题；Design finalization 的安全诊断码也会进入错误详情与可展开 TUI，而不再只显示统一失败标题。
+随包提供的 [单任务计划示例](config/plan-draft.example.json) 展示精简 PlanDraft 和结构化 Gate A 合同；按实际仓库替换示例中的路径、证据与验证命令。
+草稿可省略 `tracking`、阶段 `verificationInputs` 和现有测试的 `disposition`，编译器按任务、验证契约和已声明产物生成这些机械字段；已批准的 `changeContract` 由控制面继承，无需重复填写。
+原子验证命令可放在 `verificationDefinitions` 中，用 `{ "use": "name" }` 在各验证义务引用；Red 的 `expectedFailure` 仍需明确填写。
+内联原子验证也可省略 `id` / `classification`；任务公共 `read` 与阶段读取合并，固定 baseline/repair 保护字段由编译器补齐，恢复次数和 AGENTS 影响仍需作者明确选择。
+[多任务示例](config/plan-draft.multiple-tasks.example.json) 保留显式依赖和产物 producer；ordered steps 仍使用完整内联合同。
+`validate-plan-draft` 返回有界的最终权限、验证用途、恢复次数和实际派生来源摘要；摘要不充当批准或封存证明。
+推导有歧义时拒绝编译，显式错误仍会报错；代码不会据此扩大路径、依赖或产物权限，封存计划保持完整、严格的结构。
+预检集中返回不同任务的独立结构错误，并提供字段位置、期望/实际路径和修正提示；工具反馈与 TUI 共用有界脱敏投影。
 Design status 将可调用的 `legalOperations` 与顶层 `packetActions` 分开；显式退出只能发送 `{"action":"finish"}`，不能伪装成 `operation: "finish"`。
 
 普通 artifact、错误 Red、transport、environment、stale、conflict、baseline 与验证失败都留在同一 Implement run 中恢复。
@@ -56,9 +64,9 @@ Design 完成主要决策后，Implement 中的新实施选择默认由父模型
 增量修订或用户主动选择的 Design 完成后，新的 Implement 上下文都能本地发现并校验交付，继续同一 run。
 
 交互式 TUI、print、JSON 和 RPC 都保留 durable semantic state；父模型有自动继续动作时，TUI 显示 recovering，不把内部待编译状态显示成用户待确认，也不提示用户手动 resume。
-queued、connecting、waiting-first-response、running、validating、retrying、verifying、paused、approval-needed、applying 和 recovering 都不是完成；operation-cancelled 表示本次操作取消但 run 仍可恢复，discarded 与 rejected 是非成功终态。
+queued、preparing、waiting-first-response、running、validating、retrying、verifying、paused、approval-needed、applying 和 recovering 都不是完成；operation-cancelled 表示本次操作取消但 run 仍可恢复，discarded 与 rejected 是非成功终态。
 只有最终 apply 与 post-apply verification 提交后的 `completed` 才显示成功。
-Design finalization、Implement 终态、Diagnose/显式 `finish` 或 session shutdown 会清除 active stage 与 parent bridge；Design 还会恢复进入前的精确父工具集合，其他阶段只撤下 `abel_dispatch`。
+Design finalization、Implement 终态、Diagnose/显式 `finish` 或 session shutdown 会清除 active stage 与私有子执行状态；Design 还会恢复进入前的精确父工具集合，其他阶段只撤下 `abel_dispatch`。
 Gate 等待和可恢复暂停保持激活以接收直接后续操作。
 
 私有 journal、artifact 与 change workspace 位于 consumer repository 之外的 owner-private state root。
@@ -69,8 +77,15 @@ Gate 等待和可恢复暂停保持激活以接收直接后续操作。
 
 扩展入口负责宿主生命周期与服务装配，交付加载和验证适配器分别由独立模块负责。
 工作流入口保留原有导出，状态机仍是唯一状态转换权威；执行服务通过类型化接口返回事实。
+状态机读取运行与预算事实，独立的 `workflow-status.ts` 纯函数生成状态、阻塞批次及继续动作，不访问存储或启动执行。
+普通阶段和修复阶段共用 `candidate-artifact.ts` 校验、封存候选；恢复旧候选仍按当前批准路径检查，调用方分别处理普通越界和修复权限扩展。
 测试检查模块依赖无环，以及子会话无法导入主工作区应用和 run 状态权限。
 
+Implement 父模型属于受信任的宿主编排层，保留原有工具以调查和修复已授权的环境问题；产品修改应通过控制面完成。
+子 Worker 和控制面执行路径有代码边界；父模型直接使用宿主工具不经过这些路径，阶段指令约束其行为。
+只有 Design 会收紧父工具。
+结构化验收固定验证义务，目标是否被充分覆盖仍取决于验收测试的质量。
+
 新身份统一采用 UTF-16 码元排序，不依赖宿主语言环境。
 旧 workspace manifest 按已保存的条目顺序验证原身份；已绑定批准证明的历史交付可保留原有集合顺序，仍需验证原始字节哈希和私有批准事实。
 普通交付解析保持严格规范化。
@@ -117,7 +132,8 @@ npm install -g @abelxiaoxing/cadence
 - **本地包目录（local package directory）** — 将 Pi 指向本仓库的绝对路径或相对路径（absolute or relative path，例如 `./cadence`），Pi 发现同样的资源。
 - **已安装 tarball 目录（installed tarball directory）** — 运行 `bun pm pack --destination <tmp>` 生成真实 tarball（.tgz），安装或解压到隔离目录后指向该目录；tarball 文件本身永远不会被当作本地包传给 Pi。
 
-本包要求 Node.js `>=22.13.0`，以使用稳定可用的内置 `node:sqlite`；不从参考仓库源安装，也不维护 Pi 主机版本兼容矩阵。
+本包要求 Node.js `>=22.13.0`，以使用稳定可用的内置 `node:sqlite`；不从参考仓库源安装。
+Pi 相关 peer 依赖使用 `*`，不限制宿主 Pi 的版本；开发依赖的固定版本仅用于仓库构建与测试。
 
 ## OpenSpec 启动与平台范围
 
@@ -169,6 +185,20 @@ bun run test:target test/openspec-cli.test.ts test/design-delivery.integration.t
 inherited route 声明的能力会与当前父模型真实能力取交集，不能通过夸大的配置绕过 admission。
 Implement 保留 16,000 context / 8,000 output 的最低容量要求；任务复杂度估算只影响自动选择的优先级，满足最低要求的 route 仍可作为 fallback 或显式 rebind。
 transport 与 malformed structural result 都只在声明的 route 集合内做有界 failover；优先选择达到容量估算的 route，同一优先级保持声明顺序，只有一个 route 时会原地重试一次。
+
+子任务由 Cadence 自己维护模型／工具循环，通过 `pi-ai` 公开接口请求完整响应，不创建 Pi `AgentSession`、覆盖 Agent 钩子或扫描宿主事件历史决定终止。
+模型流事件只用于进度与完整响应适配；只有正常完整响应的工具调用可以执行，错误、截断或取消不得提交产物。
+继承路由快照当前有效 Provider、模型及每次 attempt 新鲜解析的认证，不修改父 Provider 注册表，也不要求父请求预热。
+Provider 自身的流式实现和模型配置仍然生效；宿主 session 的 `before_provider_request`／`onPayload` 不会自动移植到子任务。
+需要通用请求定制时，应使用明确配置的 Provider 实现或 custom route；子请求使用自己的 `low` reasoning 请求，最终由模型映射解释。
+输出上限沿用声明的模型／Provider 合同，不再隐式删除 Responses 的 `max_output_tokens`。
+
+子 Agent 的 evidence 提交使用精简草稿：工具绑定省略的身份，补齐未提供的提示性字段，然后按严格内部结果合同校验。
+引用、约束、风险和未决问题仍须显式提供；省略提示不等于已验证的“没有影响”，也不增加授权。
+一次被拒绝的提交可在原会话纠正一次；只有正常以文字结束且从未提交时，才额外提示补交一次。
+补交沿用原工具权限、总时限、取消和用量统计，不重新调查，不把文字直接当作交付；传输错误、截断和取消不触发补交。
+合法提交可附带简短文字，但接受后不得重复提交或覆盖结果。
+
 相同 pause 的安全指纹与重复计数通过 `status` 暴露，凭据、URL 和原始输出仍保持私有。
 `rebind` 只能选择已获 policy 授权且能力匹配的 route，不会扩大任务边界。
 
@@ -287,7 +317,7 @@ Vitest 的 JSON 报告与普通 stdout/stderr 分离，日志超过捕获预算
 
 ```sh
 bun run eval:workflow
-bun run eval:workflow --live --scenario small-fix --output /tmp/cadence-evaluation.json
+bun run eval:workflow --live --scenario small-fix --output /tmp/cadence-evaluation.json --progress-output /tmp/cadence-progress.json
 bun run eval:workflow --live --scenario multiple-tasks
 bun run eval:workflow --live --scenario restart-recovery
 bun run eval:workflow --live --scenario missing-capability
@@ -295,6 +325,30 @@ bun run eval:workflow --live --scenario missing-capability
 
 `--live` 使用 Pi 当前配置的模型，也可传 `--model provider/model`；需要可用的 Provider、OpenSpec 和 Linux Bubblewrap。
 报告记录完成状态、用户介入、重复修订、耗时及宿主报告的 token/成本；最终行为另由隔离 oracle 检查。
+`trace` 提供有界的工具/操作计数、安全错误分类、阶段里程碑、重复提交、工具活动区间并集和消息 usage；截止时冻结快照，未到达的里程碑为 null。
+非工具时间不是纯模型计算时间。
+`configurations` 记录实际模型、reasoning 和非秘密路由指纹，`sourceSha256` / `sourceUnchanged` 标识工作树版本和运行期间是否变化。
+可选 `--progress-output` 每 15 秒写一份有界进度快照；无原始参数、提示或响应正文。
+源码读取统计只覆盖明确的成功 read，未分类访问不能当作零。
+`modelFailureDiagnostics` 最多保留 16 条脱敏错误分类及可从 SDK 错误文字识别的 HTTP 状态；无法识别时明确记录 `unclassified` / `null`。
+`autoRetries` 与 `retryDelayMs` 单独记录宿主自动重试次数和计划退避时间，避免把同一请求的多次失败误读为独立故障。
 模型服务不可用、阶段停滞、取消与 Design 完成都不会被计为 Implement 成功。
 缺少能力的场景用于观察保留状态，不能把任意停滞当作正确恢复。
 无原始对话落盘。
+
+## 内部架构与请求预算
+
+`PlanDraft` 只表示作者输入，`ImplementPlan` 独立定义完整执行合同；命名引用和默认值仅在编译入口展开，旧封存计划不被重写。
+PlanDraft 的命名或内联 package-script 可以省略 `command`，编译器从 manifest 绑定脚本原文后生成身份；显式不匹配仍拒绝，Gate A 的完整验证合同不被替换。
+作者在 `tasks.md` 保留任务 checkbox、目标和精确 Scenario 归属；`compile-plan` 生成独立的验证 ID 绑定区域，保留作者正文，并在两份文件安装成功后记录编译。
+旧封存交付仍可读取；缺失作者证据或篡改绑定不能通过追溯检查。
+`workflow-state-machine` 继续独占 run/task 状态转换与预算事务；调度、恢复策略和交付差异比较使用只接受事实的内部函数。
+`durable-workflow` 管理资源生命周期和交付重放，`phase-execution` 执行候选阶段，`change-verification` 管理基线及变更验证；阶段服务没有主工作区 apply 能力。
+Pi 入口负责激活、工具交互、模型能力投影和展示；`package-workflow` 组合控制服务，`package-candidate` 执行受限候选请求，核心不接收 Pi 会话事件或完整宿主上下文。
+
+每次模型请求独立计时：90 秒内没有非空文本、思考或工具参数增量时返回 `first-progress-timeout`；已有进展后空闲 180 秒返回 `stream-idle-timeout`。
+HTTP 响应头用于观察，不刷新进展预算；完整终态可直接结束请求，本地工具执行间隙不计入流空闲，下一轮重新计时。
+Broker 保留每 attempt 20 分钟总上限和原有有限重试；该上限不代表跨重试的 workflow 总时限。
+新请求不再发出 `connect-timeout`，也不声称观测到了 TCP/TLS 建连；旧超时事实和 `connecting` 展示仍可读取。
+子执行器限制为 64 轮、4 MiB 序列化上下文，不自动压缩或静默截断证据；达到限制返回执行限制失败，候选任务暂停并要求拆分，不作为端点故障重试。
+取消后的用量结算最多等待 250 ms，迟到结果不能进入候选提交或主工作区应用。

package/agents/design-explorer.md

@@ -5,18 +5,24 @@ You are a package-owned read-only professional Agent for the `abel-design` stage
 You explore the repository ONLY through the scoped `read`, `grep`, `find`, and `ls` tools.
 You have no shell, edit, Git, network, or validation capability, and you never create or modify files.
 
-Call `abel_submit_result` exactly once with one structured object.
-Do not emit a second submit, prose, or fields outside the declared Design evidence schema.
-The object contains the matching `id`, `role: "design-explorer"`, `kind: "evidence"`, and:
+Submit one accepted final result through `abel_submit_result`; brief accompanying text is harmless.
+If the tool rejects your submission, use its error to correct once in this same disposable session.
+A second rejected structural submission ends the session.
+Never submit again after acceptance.
 
-- `packet_id` (exactly equal to the request id), `module_name`, `scope`, and
-  `files_read`;
-- `evidence`, where every item has `claim`, `path`, `line_start`, and
-  `line_end`;
-- `existing_structures`, `existing_conventions`, `constraints_discovered`,
-  `open_questions`, and `dependencies`;
-- `write_set_hints`, `validation_hints`, `agents_impact_hints`, `risks`, and
-  `success_criteria_hints`.
+Required draft fields:
+
+- `module_name`: a canonical workspace-relative module path or slug;
+- `scope` and `files_read`: unique workspace-relative paths, with nonempty scope;
+- `evidence`: cited claims with `claim`, `path`, positive `line_start`, and inclusive `line_end >= line_start`;
+- `constraints_discovered`, `open_questions`, and `risks`: explicit arrays (use `[]` only when there are none).
+
+Omit `id`, `packet_id`, `role`, and `kind`: the trusted tool binds them to this packet.
+If supplied, they must match the declared identity.
+Optional advisory arrays are `existing_structures`, `existing_conventions`, `dependencies`, `write_set_hints`, `validation_hints`, `agents_impact_hints`, and `success_criteria_hints`.
+Omission means no advisory information supplied, not a verified absence or authorization.
+Do not invent evidence or drop unresolved questions to satisfy the format.
+Do not add undeclared fields.
 
 You never return raw transcripts, hidden reasoning, or tool-call history.
 A packet is complete only when every claim is cited and every blocking question is explicit.

package/agents/diagnosis-worker.md

@@ -6,8 +6,16 @@ You collect evidence and falsify candidate root causes ONLY through the scoped `
 You never write, execute commands, or run validation.
 The fixed algorithm order is reproduce, falsify, failing-regression, then minimum-repair.
 
-An evidence packet submits exactly one compact structured evidence object through `abel_submit_result` with the reported symptoms, candidate causes, confirming or refuting citations, and a supported root-cause conclusion or an explicit evidence gap.
+An evidence packet submits one accepted compact structured evidence object through `abel_submit_result` with the reported symptoms, candidate causes, confirming or refuting citations, and a supported root-cause conclusion or an explicit evidence gap.
 Only the parent may claim that command-based reproduction succeeded.
+For evidence, omit `id`, `role`, and `kind`; the tool binds them.
+Keep `conclusions`, `citations`, `constraints`, `risks`, and `blockingQuestions` explicit.
+`dependencies` and `hints` may be omitted when no advisory information is supplied.
+If supplying hints, use `writeSet`, `verification`, and `agentsImpact` (`none`, `update-existing`, `create-index`, or `remove-index`).
+Omission never proves the absence of an impact.
+After a rejected evidence or patch submission, use the error to correct once in this same disposable session.
+A second rejected structural submission ends the session.
+Never submit again after acceptance.
 
 A candidate packet submits one complete structured patch: either the failing regression or, after the parent has verified that regression, the minimum repair.
 Use ordered `replace`, `rewrite`, `create`, and `delete` operations inside the approved write set.

package/agents/implementation-worker.md

@@ -8,7 +8,7 @@ You never write the workspace, run validation, or execute commands.
 Your output is one complete structured patch delivered through `abel_submit_result`:
 
 - copy the attempt-bound `candidateId` from the phase contract;
-- submit exactly one `candidate-patch` call containing that `candidateId` and
+- submit one accepted `candidate-patch` call containing that `candidateId` and
   an ordered `operations` array;
 - use `replace` with exact uniquely occurring `oldText` and the intended
   `newText` for compact edits, `rewrite` for a complete existing-file body,
@@ -32,6 +32,8 @@ Your output is one complete structured patch delivered through `abel_submit_resu
 
 If the submit tool rejects malformed operations, use its concrete error to correct the submission once in this same disposable session; a second rejected structural submission ends the session.
 
+Never submit again after acceptance; brief accompanying text is harmless.
+
 If the configured candidate limit is exceeded, the submit tool returns terminal typed `result-limit` with `limitBytes`; never continue with a partial or truncated patch.
 The parent alone reviews, applies, and validates the generated diff.
 You report candidate facts only and never claim execution success, select parent recovery, choose a workflow stage, or control the task DAG.

package/config/plan-draft.example.json

@@ -0,0 +1,122 @@
+{
+  "changeContract": {
+    "goal": "Fix integer addition while preserving the existing zero case",
+    "acceptance": [
+      {
+        "id": "integer-addition",
+        "statement": "add(2, 3) returns 5 and add(0, 0) remains 0",
+        "verification": {
+          "kind": "static-check",
+          "id": "add-acceptance",
+          "runner": {
+            "kind": "node",
+            "script": "test/add.test.mjs"
+          },
+          "args": [],
+          "classification": "expected-green"
+        }
+      }
+    ],
+    "constraints": [
+      {
+        "id": "no-dependencies",
+        "statement": "Keep the existing function signature and add no dependencies"
+      }
+    ],
+    "policy": {
+      "writeRoots": ["src/add.mjs", "test/add.test.mjs"],
+      "dependencies": [],
+      "verificationModes": ["behavior"]
+    }
+  },
+  "changeId": "fix-integer-addition",
+  "tasks": [
+    {
+      "taskId": "fix-add",
+      "dependsOn": [],
+      "objective": "Preserve the existing zero assertion; add a failing positive-integer assertion with [ADD:positive-integers], then fix addition",
+      "context": {
+        "agents": "No AGENTS.md in this example consumer; use the actual repository instructions when adapting",
+        "contract": "Gate A integer-addition acceptance; only src/add.mjs and test/add.test.mjs may change"
+      },
+      "roots": ["."],
+      "phases": {
+        "red": {
+          "write": ["test/add.test.mjs"],
+          "delete": [],
+          "verification": {
+            "use": "add-test",
+            "expectedFailure": "[ADD:positive-integers]"
+          }
+        },
+        "green": {
+          "write": ["src/add.mjs"],
+          "delete": [],
+          "verification": {
+            "use": "add-test"
+          }
+        }
+      },
+      "affectedVerification": {
+        "use": "add-test"
+      },
+      "repairVerification": {
+        "use": "add-test"
+      },
+      "scheduling": {
+        "conflicts": [],
+        "resources": []
+      },
+      "agents": {
+        "impact": "none"
+      },
+      "approvedDependencies": [],
+      "impactClosure": {
+        "changedSurfaces": ["none"],
+        "searchEvidence": [
+          "Example consumer: src/add.mjs exports add; test/add.test.mjs is its only caller and existing test"
+        ],
+        "relatedTests": [
+          {
+            "path": "test/add.test.mjs",
+            "evidence": "Existing add(0, 0) assertion passes; preserve it while adding the positive-integer regression"
+          }
+        ],
+        "affectedSuite": ["test/add.test.mjs"]
+      },
+      "read": ["package.json", "src/add.mjs", "test/add.test.mjs"]
+    }
+  ],
+  "outputs": [],
+  "verification": {
+    "baseline": {
+      "fullSuite": {
+        "use": "add-test"
+      }
+    },
+    "change": {
+      "fullSuite": {
+        "use": "add-test"
+      },
+      "postApply": {
+        "use": "add-test"
+      }
+    },
+    "artifactCorrection": {
+      "maxAttempts": 2
+    },
+    "repair": {
+      "maxAttempts": 1
+    }
+  },
+  "verificationDefinitions": {
+    "add-test": {
+      "kind": "static-check",
+      "runner": {
+        "kind": "node",
+        "script": "test/add.test.mjs"
+      },
+      "args": []
+    }
+  }
+}

package/config/plan-draft.multiple-tasks.example.json

@@ -0,0 +1,345 @@
+{
+  "changeContract": {
+    "goal": "Fix integer addition and doubling through add",
+    "acceptance": [
+      {
+        "id": "integer-addition",
+        "statement": "add(2, 3) returns 5 and add(0, 0) remains 0",
+        "verification": {
+          "kind": "static-check",
+          "id": "add-acceptance",
+          "runner": {
+            "kind": "node",
+            "script": "test/add.test.mjs"
+          },
+          "args": [],
+          "classification": "expected-green"
+        }
+      },
+      {
+        "id": "integer-double",
+        "statement": "double(3) returns 6 via add(3, 3)",
+        "verification": {
+          "kind": "steps",
+          "id": "double-green",
+          "classification": "expected-green",
+          "steps": [
+            {
+              "kind": "static-check",
+              "id": "double-green-add",
+              "runner": {
+                "kind": "node",
+                "script": "test/add.test.mjs"
+              },
+              "args": [],
+              "classification": "expected-green"
+            },
+            {
+              "kind": "static-check",
+              "id": "double-green-own",
+              "runner": {
+                "kind": "node",
+                "script": "test/double.test.mjs"
+              },
+              "args": [],
+              "classification": "expected-green"
+            }
+          ]
+        }
+      }
+    ],
+    "constraints": [
+      {
+        "id": "no-dependencies",
+        "statement": "Keep the existing function signature and add no dependencies"
+      }
+    ],
+    "policy": {
+      "writeRoots": [
+        "src/add.mjs",
+        "test/add.test.mjs",
+        "src/double.mjs",
+        "test/double.test.mjs"
+      ],
+      "dependencies": [],
+      "verificationModes": ["behavior"]
+    }
+  },
+  "changeId": "fix-add-and-double",
+  "tasks": [
+    {
+      "taskId": "fix-add",
+      "dependsOn": [],
+      "objective": "Preserve the existing zero assertion; add a failing positive-integer assertion with [ADD:positive-integers], then fix addition",
+      "context": {
+        "agents": "No AGENTS.md in this example consumer; use the actual repository instructions when adapting",
+        "contract": "Gate A integer-addition acceptance; only src/add.mjs and test/add.test.mjs may change"
+      },
+      "roots": ["."],
+      "phases": {
+        "red": {
+          "write": ["test/add.test.mjs"],
+          "delete": [],
+          "verification": {
+            "use": "add-test",
+            "expectedFailure": "[ADD:positive-integers]"
+          }
+        },
+        "green": {
+          "write": ["src/add.mjs"],
+          "delete": [],
+          "verification": {
+            "use": "add-test"
+          }
+        }
+      },
+      "affectedVerification": {
+        "use": "add-test"
+      },
+      "repairVerification": {
+        "use": "add-test"
+      },
+      "scheduling": {
+        "conflicts": [],
+        "resources": []
+      },
+      "agents": {
+        "impact": "none"
+      },
+      "approvedDependencies": [],
+      "impactClosure": {
+        "changedSurfaces": ["none"],
+        "searchEvidence": [
+          "Example consumer: src/add.mjs exports add; test/add.test.mjs is its only caller and existing test"
+        ],
+        "relatedTests": [
+          {
+            "path": "test/add.test.mjs",
+            "evidence": "Existing add(0, 0) assertion passes; preserve it while adding the positive-integer regression"
+          }
+        ],
+        "affectedSuite": ["test/add.test.mjs"]
+      },
+      "read": ["package.json", "src/add.mjs", "test/add.test.mjs"]
+    },
+    {
+      "taskId": "fix-double",
+      "dependsOn": ["fix-add"],
+      "objective": "Keep double(0) passing; add the [DOUBLE:positive-integers] failure then use add(value, value)",
+      "context": {
+        "agents": "No AGENTS.md in this example consumer; use the actual repository instructions when adapting",
+        "contract": "Gate A integer-double acceptance; depends on corrected addition"
+      },
+      "roots": ["."],
+      "phases": {
+        "red": {
+          "write": ["test/double.test.mjs"],
+          "delete": [],
+          "verification": {
+            "kind": "steps",
+            "id": "double-red",
+            "classification": "expected-red",
+            "steps": [
+              {
+                "kind": "static-check",
+                "id": "double-red-add",
+                "runner": {
+                  "kind": "node",
+                  "script": "test/add.test.mjs"
+                },
+                "args": [],
+                "classification": "expected-green"
+              },
+              {
+                "kind": "static-check",
+                "id": "double-red-own",
+                "runner": {
+                  "kind": "node",
+                  "script": "test/double.test.mjs"
+                },
+                "args": [],
+                "classification": "expected-red",
+                "expectedFailure": "[DOUBLE:positive-integers]"
+              }
+            ]
+          }
+        },
+        "green": {
+          "write": ["src/double.mjs"],
+          "delete": [],
+          "verification": {
+            "kind": "steps",
+            "id": "double-green",
+            "classification": "expected-green",
+            "steps": [
+              {
+                "kind": "static-check",
+                "id": "double-green-add",
+                "runner": {
+                  "kind": "node",
+                  "script": "test/add.test.mjs"
+                },
+                "args": [],
+                "classification": "expected-green"
+              },
+              {
+                "kind": "static-check",
+                "id": "double-green-own",
+                "runner": {
+                  "kind": "node",
+                  "script": "test/double.test.mjs"
+                },
+                "args": [],
+                "classification": "expected-green"
+              }
+            ]
+          }
+        }
+      },
+      "affectedVerification": {
+        "kind": "steps",
+        "id": "double-affectedVerification",
+        "classification": "expected-green",
+        "steps": [
+          {
+            "kind": "static-check",
+            "id": "double-green-add-affectedVerification",
+            "runner": {
+              "kind": "node",
+              "script": "test/add.test.mjs"
+            },
+            "args": [],
+            "classification": "expected-green"
+          },
+          {
+            "kind": "static-check",
+            "id": "double-green-own-affectedVerification",
+            "runner": {
+              "kind": "node",
+              "script": "test/double.test.mjs"
+            },
+            "args": [],
+            "classification": "expected-green"
+          }
+        ]
+      },
+      "repairVerification": {
+        "kind": "steps",
+        "id": "double-repairVerification",
+        "classification": "expected-green",
+        "steps": [
+          {
+            "kind": "static-check",
+            "id": "double-green-add-repairVerification",
+            "runner": {
+              "kind": "node",
+              "script": "test/add.test.mjs"
+            },
+            "args": [],
+            "classification": "expected-green"
+          },
+          {
+            "kind": "static-check",
+            "id": "double-green-own-repairVerification",
+            "runner": {
+              "kind": "node",
+              "script": "test/double.test.mjs"
+            },
+            "args": [],
+            "classification": "expected-green"
+          }
+        ]
+      },
+      "scheduling": {
+        "conflicts": [],
+        "resources": []
+      },
+      "agents": {
+        "impact": "none"
+      },
+      "approvedDependencies": [],
+      "impactClosure": {
+        "changedSurfaces": ["none"],
+        "searchEvidence": [
+          "Example consumer: double imports add; each has an existing zero assertion; all.test.mjs imports both test files"
+        ],
+        "relatedTests": [
+          {
+            "path": "test/add.test.mjs",
+            "evidence": "Producer task owns the addition regression, checked before the double regression"
+          },
+          {
+            "path": "test/double.test.mjs",
+            "evidence": "Preserve the existing double(0) assertion while adding double(3)"
+          }
+        ],
+        "affectedSuite": ["test/add.test.mjs", "test/double.test.mjs"]
+      },
+      "read": [
+        "package.json",
+        "src/add.mjs",
+        "test/add.test.mjs",
+        "src/double.mjs",
+        "test/double.test.mjs"
+      ]
+    }
+  ],
+  "outputs": [
+    {
+      "id": "add-regression",
+      "path": "test/add.test.mjs",
+      "producer": {
+        "taskId": "fix-add",
+        "phase": "red"
+      },
+      "postcondition": "regular-file"
+    }
+  ],
+  "verification": {
+    "baseline": {
+      "fullSuite": {
+        "use": "full-suite"
+      }
+    },
+    "change": {
+      "fullSuite": {
+        "use": "full-suite"
+      },
+      "postApply": {
+        "use": "full-suite"
+      }
+    },
+    "artifactCorrection": {
+      "maxAttempts": 2
+    },
+    "repair": {
+      "maxAttempts": 1
+    }
+  },
+  "verificationDefinitions": {
+    "add-test": {
+      "kind": "static-check",
+      "runner": {
+        "kind": "node",
+        "script": "test/add.test.mjs"
+      },
+      "args": []
+    },
+    "full-suite": {
+      "kind": "static-check",
+      "runner": {
+        "kind": "node",
+        "script": "test/all.test.mjs"
+      },
+      "args": []
+    },
+    "double-test": {
+      "kind": "static-check",
+      "runner": {
+        "kind": "node",
+        "script": "test/double.test.mjs"
+      },
+      "args": []
+    }
+  }
+}