agent-exec transport c4e7bcc1003fb3b60d968183cc9674cd/2e81d2e9845e

AGENT-EXECUTION.md

@@ -18,8 +18,8 @@ increment host fallback authorizations. Safe exhaustion without an emergency
 candidate ends `unavailable` or `blocked`, and that result does not authorize
 the host to write the unit. The legacy value `all-unavailable` remains
 parseable and likewise grants no host write authority. Cursor uses
-`cursor-agent-stable`. `model: auto` lets each CLI use its configured default
-model; PI therefore uses the operator-selected MiniMax default. The
+`cursor-agent-stable`. `model: auto` lets a CLI use its configured default
+model; any other value is passed to that CLI as its model selector. The
 `integration` block declares the host-exempt paths and the transport, admission,
 and cleanup entry points.
 
@@ -60,7 +60,7 @@ profiles:
   pi:
     adapter: pi
     executable: pi
-    model: auto
+    model: minimax-cn/MiniMax-M3
     enabled: true
 
   codex:
@@ -78,3 +78,38 @@ integration:
   cleanup:
     command: make task-finish
 ```
+
+## Model preferences (project choice)
+
+Generic rules for bounded packets, context budgeting, phase handoff and
+requested-versus-effective model identity live in the shared
+[standards/agent-execution.md](deps/development-guidelines/standards/agent-execution.md)
+and [workflows/handoff.md](deps/development-guidelines/workflows/handoff.md);
+project procedure is in
+[policy/delivery.md](policy/delivery.md#有界单元包上下文与阶段交接). This
+section records only this project's choices.
+
+- Planning, design and implementation plans: Codex or Claude when available,
+  in the host or a fresh CLI session. This is a preference, not a routing role.
+- Bounded implementation, repair and test units: the existing normal
+  candidates in the configured role order above. Requested models:
+  - PI: `minimax-cn/MiniMax-M3`, passed as `--model provider/id`. PI
+    actual events have reported `minimax-cn` / `MiniMax-M3` and the local
+    offline catalog lists that profile; each run still records the effective
+    model from that run's events.
+  - Cursor: `auto` (Cursor Auto).
+  - agy: Gemini 3.8 Flash is requested, but its exact selectable model ID and
+    effective identity are `NOT_VERIFIED` (the `agy` models query needs a
+    login that was not available). The profile therefore stays `model: auto`
+    until an observed model ID exists; do not claim Gemini 3.8 Flash is active.
+- Assigning a different candidate or model to a unit needs a concrete,
+  task-specific reason recorded in the task record, and stays within the
+  existing roles and candidate lists. It is not a silent permission or routing
+  change. Role lists, attempt limits, timeouts, output limits, the single
+  Codex emergency attempt, stage isolation and `fallback_to_host: never` are
+  unchanged by these preferences.
+- A requested model is prompt/config intent. Only the model reported in a
+  run's actual event envelope counts as effective; when the CLI does not report
+  it, record `not-observable`. Unknown context capacity, a missing agy login
+  or missing automatic compaction are recorded facts, not global fail-closed
+  conditions.

policy/delivery.md

@@ -154,6 +154,30 @@ host 可以直接编写并提交的豁免只有 `.task/**`、`docs/superpowers/s
 - 会话隔离不要求为简单任务增加新平台或复杂编排。平凡文档/配置任务仍按最小闭环执行；需要独立 Review
   时只额外启动一个最小、一次性的评审 session，不复用实施 session，也不扩展任务范围。
 
+### 有界单元包、上下文与阶段交接
+
+通用规则见共享 [standards/agent-execution.md](../deps/development-guidelines/standards/agent-execution.md)
+与 [workflows/handoff.md](../deps/development-guidelines/workflows/handoff.md)；本节只写本仓做法，模型偏好见
+[AGENT-EXECUTION.md](../AGENT-EXECUTION.md#model-preferences-project-choice)。
+
+- **单元包**：用现有单元包 JSON 的 `goal`/`constraints`/`verification` 写清：一个结果及验收条件、允许路径与
+  须遵守的契约、非目标、前置提交 SHA、最小必读清单、直接验证命令、上下文容量及其来源（未知时按保守的
+  100K 假设）和为执行预留的余量。不新增 schema、运行时字段或控制器。
+- **容量**：100K/200K/300K/500K 只是规划参考，不是对任何 provider 能力的声明；容量更大不授权更宽的任务，
+  单元仍按一个结果拆分。容量未知不构成失败条件，也不做自动容量强制。
+- **阶段交接**：长任务每完成一个里程碑，host 自动在现有任务记录中写一份简短的持久交接：目标、状态、关键决定、
+  文件与 SHA、命令与结果、阻塞项、下一步和来源链接；下一个 CLI 阶段或单元以全新会话启动，只读取交接与最小
+  必读材料，不为交接等待人工。
+- **上下文压力**：provider 支持时，只在阶段完成后使用其上下文压缩；没有自动压缩不阻塞工作。普通上下文压力用
+  交接加 host 在既有授权内重新打包单元处理；只有真实扩围才按既有 `scope-expansion` fail-stop。
+- **如实区分**：请求模型与实际事件报告的有效模型；可观察的 session 身份与 `not-observable`；worker 局部自检
+  （SELF）与 host 集成验证、独立 Review、精确 HEAD CI（GATE）。
+- **验证分工**：worker 只跑直接检查；host 在最终候选上运行一次所需的集成套件。涉及安全、契约或部署的改动，
+  须写明正向、负向与回归用例。
+- **单写者与强制边界**：同一单元同一时刻只有一个写者，host 在运行期间不改规范检出。记录时区分三类约束：
+  提示词指导（单元包内容、容量、交接、模型偏好）、观察到的行为（事件报告的模型与 session），以及工具强制
+  （配置校验、尝试上限、超时、transport/admit）。本节不新增门禁，不放宽凭据、范围、未知副作用与 CI 边界。
+
 ### 两级分支交付模型与分流检查
 
 本仓采用两级业务交付模型：