@fitlab-ai/agent-infra 0.9.1 → 0.9.2

package/templates/.agents/skills/upgrade-dependency/SKILL.en.md

@@ -63,13 +63,13 @@ Report:
 
 Suggest next step:
 
-> **IMPORTANT**: All TUI command formats listed below must be output in full. Do not show only the format for the current AI agent. If `.agents/.airc.json` configures custom TUIs (via `customTUIs`), read each tool's `name` and `invoke`, then add the matching command line in the same format (`${skillName}` becomes the skill name and `${projectName}` becomes the project name).
+> Before rendering next steps, read `.agents/rules/next-step-output.md`, invoke the shared helper only for the selected scenario, and insert its stdout at `{next-step-commands}`.
+
+Populate `{next-step-commands}` for this scenario by running `agent-infra-internal agent-client next-steps --skill commit`.
 
 ```
 Next step - commit changes:
-  - Claude Code / OpenCode: /commit
-  - Gemini CLI: /{{project}}:commit
-  - Codex CLI: $commit
+{next-step-commands}
 ```
 
 ## Notes

package/templates/.agents/skills/upgrade-dependency/SKILL.zh-CN.md

@@ -63,13 +63,13 @@ description: >
 
 建议下一步：
 
-> **重要**：以下「下一步」中列出的所有 TUI 命令格式必须完整输出，不要只展示当前 AI 代理对应的格式。如果 `.agents/.airc.json` 中配置了自定义 TUI（`customTUIs`），读取每个工具的 `name` 和 `invoke`，按同样格式补充对应命令行（`${skillName}` 替换为技能名，`${projectName}` 替换为项目名）。
+> 渲染下一步前先读取 `.agents/rules/next-step-output.md`，仅为已选场景调用统一 helper，并将 stdout 填入 `{next-step-commands}`。
+
+使用 `agent-infra-internal agent-client next-steps --skill commit` 生成本场景的 `{next-step-commands}`。
 
 ```
 下一步 - 提交代码：
-  - Claude Code / OpenCode：/commit
-  - Gemini CLI：/{{project}}:commit
-  - Codex CLI：$commit
+{next-step-commands}
 ```
 
 ## 注意事项

package/templates/.agents/skills/watch-pr/SKILL.en.md

@@ -101,27 +101,27 @@ Keep the gate output in your reply as the verification evidence. Without current
 
 > On the task-anchored path, execute this step only after the gate passes.
 
-> **IMPORTANT**: All TUI command formats listed below must be output in full. Do not show only the format for the current AI agent. If `.agents/.airc.json` configures custom TUIs (via `customTUIs`), read each tool's `name` and `invoke`, then add the matching command line in the same format (`${skillName}` becomes the skill name and `${projectName}` becomes the project name). Before rendering the final output, read `.agents/rules/next-step-output.md` and apply both of its rules: (1) render `{task-ref}` in the "Next steps" commands as the short id `NN` (falling back to the full TASK-id when unallocated or released); (2) append the `Completed at` line as the very last line of the user-facing output (this applies to every user-facing output — success, error, and early-return paths alike, not only the success path).
+> Before rendering next steps, read `.agents/rules/next-step-output.md`, invoke the shared helper only for the selected scenario, and insert its stdout at `{next-step-commands}`.
 
 Output per scenario:
 - "All green" + task-anchored: state that all required checks passed, then render exactly one exit based on whether this run created repair commits (`{task-ref}` becomes the short id):
 
   `repairCommits.length == 0`:
 
+Populate `{next-step-commands}` for this scenario by running `agent-infra-internal agent-client next-steps --skill complete-task --task-ref {task-ref}`.
+
   ```
   Next step - Complete and archive the task:
-    - Claude Code / OpenCode: /complete-task {task-ref}
-    - Gemini CLI: /{{project}}:complete-task {task-ref}
-    - Codex CLI: $complete-task {task-ref}
+  {next-step-commands}
   ```
 
   `repairCommits.length > 0`:
 
+Populate `{next-step-commands}` for this scenario by running `agent-infra-internal agent-client next-steps --skill review-code --task-ref {task-ref}`.
+
   ```
   Next step - Re-run code review:
-    - Claude Code / OpenCode: /review-code {task-ref}
-    - Gemini CLI: /{{project}}:review-code {task-ref}
-    - Codex CLI: $review-code {task-ref}
+  {next-step-commands}
   ```
 
 - "Blocked": output only the step 4 blocker explanation; do not recommend a next-step command.
@@ -133,7 +133,7 @@ Output per scenario:
 - [ ] Self-heal limited to locatable code-layer failures, with local tests passing before push and within the fix cap
 - [ ] Task-anchored path: updated task.md and appended the Watch PR Activity Log entry
 - [ ] Task-anchored path: verification gate passed
-- [ ] Showed the user all TUI next-step command formats (green exit; the blocked exit renders no next step)
+- [ ] Rendered the selected next-step commands through the shared helper
 
 ## Stop
 

package/templates/.agents/skills/watch-pr/SKILL.zh-CN.md

@@ -101,27 +101,27 @@ agent-infra-internal task-verify {task-id} watch-pr.completed --format text
 
 > 任务锚定路径仅在校验通过后执行本步骤。
 
-> **重要**：以下「下一步」中列出的所有 TUI 命令格式必须完整输出，不要只展示当前 AI 代理对应的格式。如果 `.agents/.airc.json` 中配置了自定义 TUI（`customTUIs`），读取每个工具的 `name` 和 `invoke`，按同样格式补充对应命令行（`${skillName}` 替换为技能名，`${projectName}` 替换为项目名）。 渲染最终输出前，先读取 `.agents/rules/next-step-output.md` 并落实其两类规则：(1) 「下一步」命令把 `{task-ref}` 渲染为短号 `NN`（未分配/已释放时回退完整 TASK-id）；(2) 在面向用户输出的绝对最后一行追加 `Completed at` 收尾行（成功、错误、早退等任何面向用户输出都适用，不限于校验通过的成功态）。
+> 渲染下一步前先读取 `.agents/rules/next-step-output.md`，仅为已选场景调用统一 helper，并将 stdout 填入 `{next-step-commands}`。
 
 按场景输出：
 - 「全绿」+ 任务锚定：说明所有 required checks 已通过，并只按本轮是否产生修复 commit 渲染一个出口（`{task-ref}` 替换为短号）：
 
   `repairCommits.length == 0`：
 
+使用 `agent-infra-internal agent-client next-steps --skill complete-task --task-ref {task-ref}` 生成本场景的 `{next-step-commands}`。
+
   ```
   下一步 - 完成并归档任务：
-    - Claude Code / OpenCode：/complete-task {task-ref}
-    - Gemini CLI：/{{project}}:complete-task {task-ref}
-    - Codex CLI：$complete-task {task-ref}
+  {next-step-commands}
   ```
 
   `repairCommits.length > 0`：
 
+使用 `agent-infra-internal agent-client next-steps --skill review-code --task-ref {task-ref}` 生成本场景的 `{next-step-commands}`。
+
   ```
   下一步 - 重新代码审查：
-    - Claude Code / OpenCode：/review-code {task-ref}
-    - Gemini CLI：/{{project}}:review-code {task-ref}
-    - Codex CLI：$review-code {task-ref}
+  {next-step-commands}
   ```
 
 - 「阻塞」：仅输出步骤 4 的阻塞说明，不推荐下一步命令。
@@ -133,7 +133,7 @@ agent-infra-internal task-verify {task-id} watch-pr.completed --format text
 - [ ] 自愈仅限可定位的代码层失败，且推送前本地测试通过、未超修复上限
 - [ ] 任务锚定路径：更新了 task.md 并追加 Watch PR 的 Activity Log
 - [ ] 任务锚定路径：完成校验通过
-- [ ] 向用户展示了所有 TUI 格式的下一步命令（全绿出口；阻塞出口不渲染下一步）
+- [ ] 已通过统一 helper 渲染已选场景的下一步命令
 
 ## 停止
 

package/templates/.claude/agents/agent-infra-lifecycle-executor.en.md

@@ -0,0 +1,6 @@
+---
+name: agent-infra-lifecycle-executor
+description: Execute exactly one agent-infra lifecycle stage in a fresh context.
+---
+
+Run only the requested non-review lifecycle skill for the supplied task reference. Do not run any `review-*` skill. Preserve the existing worktree, follow the selected skill completely, and stop after its completion output.

package/templates/.claude/agents/agent-infra-lifecycle-executor.zh-CN.md

@@ -0,0 +1,6 @@
+---
+name: agent-infra-lifecycle-executor
+description: 在全新上下文中只执行一个 agent-infra 生命周期阶段。
+---
+
+只为给定任务引用运行指定的非审查生命周期技能。不得运行任何 `review-*` 技能。保留现有工作树，完整遵循所选技能，并在该技能完成输出后停止。

package/templates/.claude/agents/agent-infra-lifecycle-reviewer.en.md

@@ -0,0 +1,6 @@
+---
+name: agent-infra-lifecycle-reviewer
+description: Review exactly one agent-infra lifecycle artifact in a fresh isolated context.
+---
+
+Run only the requested `review-*` skill for the supplied task reference. Do not modify business code, do not execute implementation skills, and stop after producing and finalizing the review artifact.

package/templates/.claude/agents/agent-infra-lifecycle-reviewer.zh-CN.md

@@ -0,0 +1,6 @@
+---
+name: agent-infra-lifecycle-reviewer
+description: 在全新隔离上下文中只审查一个 agent-infra 生命周期产物。
+---
+
+只为给定任务引用运行指定的 `review-*` 技能。不得修改业务代码，不得执行实现技能，并在生成和最终化审查产物后停止。

package/templates/.claude/commands/run-task.en.md

@@ -0,0 +1,8 @@
+---
+description: "Orchestrate a task lifecycle with fresh isolated agents"
+usage: "/run-task [task-ref | --task <ref>]"
+---
+
+Read and execute the run-task skill from `.agents/skills/run-task/SKILL.md`.
+
+Follow all steps defined in the skill exactly.

package/templates/.claude/commands/run-task.zh-CN.md

@@ -0,0 +1,8 @@
+---
+description: "使用 fresh 隔离 Agent 总控任务生命周期"
+usage: "/run-task [task-ref | --task <ref>]"
+---
+
+读取并执行 `.agents/skills/run-task/SKILL.md` 中的 run-task 技能。
+
+严格按照技能中定义的所有步骤执行。

package/templates/.claude/settings.json

@@ -30,6 +30,30 @@
         ]
       }
     ],
+    "SubagentStart": [
+      {
+        "matcher": "agent-infra-lifecycle-(executor|reviewer)",
+        "hooks": [
+          {
+            "type": "command",
+            "command": "node \"$(git rev-parse --show-toplevel)/.agents/hooks/lifecycle-delegation.js\" --client claude-code",
+            "timeout": 10
+          }
+        ]
+      }
+    ],
+    "SubagentStop": [
+      {
+        "matcher": "agent-infra-lifecycle-(executor|reviewer)",
+        "hooks": [
+          {
+            "type": "command",
+            "command": "node \"$(git rev-parse --show-toplevel)/.agents/hooks/lifecycle-delegation.js\" --client claude-code",
+            "timeout": 10
+          }
+        ]
+      }
+    ],
     "StopFailure": [
       {
         "matcher": "",

package/templates/.codex/agents/agent-infra-lifecycle-executor.en.toml

@@ -0,0 +1,3 @@
+name = "agent-infra-lifecycle-executor"
+description = "Execute exactly one agent-infra lifecycle stage in a fresh thread."
+developer_instructions = "Run only the requested non-review lifecycle skill for the supplied task reference. Never run review-* skills. Stop after the stage completes."

package/templates/.codex/agents/agent-infra-lifecycle-executor.zh-CN.toml

@@ -0,0 +1,3 @@
+name = "agent-infra-lifecycle-executor"
+description = "在全新线程中只执行一个 agent-infra 生命周期阶段。"
+developer_instructions = "只为给定任务引用运行指定的非审查生命周期技能。不得运行 review-* 技能。阶段完成后停止。"

package/templates/.codex/agents/agent-infra-lifecycle-reviewer.en.toml

@@ -0,0 +1,3 @@
+name = "agent-infra-lifecycle-reviewer"
+description = "Review exactly one lifecycle artifact in a fresh isolated thread."
+developer_instructions = "Run only the requested review-* skill. Never modify business code or execute implementation skills. Stop after the review artifact is finalized."

package/templates/.codex/agents/agent-infra-lifecycle-reviewer.zh-CN.toml

@@ -0,0 +1,3 @@
+name = "agent-infra-lifecycle-reviewer"
+description = "在全新隔离线程中只审查一个生命周期产物。"
+developer_instructions = "只运行指定的 review-* 技能。不得修改业务代码或执行实现技能。审查产物最终化后停止。"

package/templates/.gemini/commands/_project_/run-task.en.toml

@@ -0,0 +1,8 @@
+description = "Orchestrate a task lifecycle with fresh isolated agents"
+prompt = """
+Run task lifecycle {{args}}.
+
+Read and execute the run-task skill from `.agents/skills/run-task/SKILL.md`.
+
+Follow all steps defined in the skill exactly.
+"""

package/templates/.gemini/commands/_project_/run-task.zh-CN.toml

@@ -0,0 +1,8 @@
+description = "使用 fresh 隔离 Agent 总控任务生命周期"
+prompt = """
+运行任务生命周期 {{args}}。
+
+读取并执行 `.agents/skills/run-task/SKILL.md` 中的 run-task 技能。
+
+严格按照技能中定义的所有步骤执行。
+"""

package/templates/.opencode/commands/run-task.en.md

@@ -0,0 +1,11 @@
+---
+description: "Orchestrate a task lifecycle with fresh isolated agents"
+agent: general
+subtask: false
+---
+
+Run task lifecycle $ARGUMENTS.
+
+Read and execute the run-task skill from `.agents/skills/run-task/SKILL.md`.
+
+Follow all steps defined in the skill exactly.

package/templates/.opencode/commands/run-task.zh-CN.md

@@ -0,0 +1,11 @@
+---
+description: "使用 fresh 隔离 Agent 总控任务生命周期"
+agent: general
+subtask: false
+---
+
+运行任务生命周期 $ARGUMENTS。
+
+读取并执行 `.agents/skills/run-task/SKILL.md` 中的 run-task 技能。
+
+严格按照技能中定义的所有步骤执行。