@fitlab-ai/agent-infra 0.9.3 → 0.9.4

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

@@ -1,17 +1,17 @@
 ---
 name: watch-pr
 description: >
-  Watch PR readiness and self-heal required-check failures or merge conflicts.
+  Watch PR readiness and self-heal any check failure or merge conflict.
   Use when a PR must be monitored until checks pass and it is explicitly mergeable.
 ---
 
 # Watch Pull Request
 
-After `create-pr`, continuously watch required checks and mergeability for the same PR head. Only passed checks plus explicit mergeability may reach success; check failures use the existing repair path, conflicts use constrained rebase healing, and unknown or unsafe states fail closed.
+After `create-pr`, continuously watch all checks and mergeability for the same PR head. Only all passed checks plus explicit mergeability may reach success; check failures use the existing repair path, conflicts use constrained rebase healing, and unknown or unsafe states fail closed.
 
 ## Behavior Boundaries / Key Rules
 
-- Only self-heal the current PR's required checks and text merge conflicts; do not add approval or repository-rule readiness dimensions.
+- Only self-heal the current PR's full check set and text merge conflicts; do not add approval or repository-rule readiness dimensions.
 - Self-heal publishes through Git workflow intents, but **affected tests must pass first**; the attempt cap and code-layer authorization remain unchanged.
 - The help exit is "produce-then-stop": end this round, output the blocker explanation, and wait for the user to trigger the next step — **never** ask mid-flow.
 - Bare numbers / `NN` / `TASK-id` arguments are always resolved as task short ids (see `.agents/rules/task-short-id.md`); a PR number is passed only via `--pr <number>` / a PR URL / omission (current branch), never reusing the bare-number syntax.
@@ -104,7 +104,7 @@ Keep the gate output in your reply as the verification evidence. Without current
 > 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:
-- `ready` + task-anchored: state that required checks passed and the current head is explicitly mergeable, then render exactly one exit based on whether this run created repair commits (`{task-ref}` becomes the short id):
+- `ready` + task-anchored: state that all checks passed and the current head is explicitly mergeable, then render exactly one exit based on whether this run created repair commits (`{task-ref}` becomes the short id):
 
   `repairCommits.length == 0`:
 

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

@@ -1,17 +1,17 @@
 ---
 name: watch-pr
 description: >
-  监控 PR readiness，并在 required checks 失败或合并冲突时自愈。
+  监控 PR readiness，并在任一 check 失败或合并冲突时自愈。
   当需要持续监控 PR 直到门禁通过且明确可合入时使用。
 ---
 
 # 监控 Pull Request
 
-在 `create-pr` 之后持续监控同一 PR head 的 required checks 与 mergeability。只有 checks 通过且平台明确可合入才进入成功出口；check 失败走既有修复，合并冲突走受限 rebase 自愈，未知或无法安全闭环时 fail-closed。
+在 `create-pr` 之后持续监控同一 PR head 的全部 checks 与 mergeability。只有全部 checks 通过且平台明确可合入才进入成功出口；check 失败走既有修复，合并冲突走受限 rebase 自愈，未知或无法安全闭环时 fail-closed。
 
 ## 行为边界 / 关键规则
 
-- 仅监控 + 自愈当前 PR 的 required checks 与文本合并冲突；不扩展到审批或其他仓库规则。
+- 仅监控 + 自愈当前 PR 的全部 checks 与文本合并冲突；不扩展到审批或其他仓库规则。
 - 自愈通过 Git workflow intent 发布修复，但**发布前必须本地跑通相关测试**；修复上限与代码层分类授权不变。
 - 求助出口是「产出后停止」语义：停止本轮、输出阻塞说明、等待用户主动触发，**不**中途提问。
 - 裸数字 / `NN` / `TASK-id` 入参一律按任务短号解析（见 `.agents/rules/task-short-id.md`）；PR 号只走 `--pr <number>` / PR URL / 省略（当前分支），不复用裸数字语法。
@@ -104,7 +104,7 @@ agent-infra-internal task-verify {task-id} watch-pr.completed --format text
 > 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：
-- `ready` + 任务锚定：说明 required checks 已通过且当前 head 明确可合入，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）：
+- `ready` + 任务锚定：说明全部 checks 已通过且当前 head 明确可合入，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`：
 

package/templates/.agents/skills/watch-pr/reference/monitor-and-heal.en.md

@@ -6,8 +6,8 @@ Platform-agnostic decision logic for `watch-pr` steps 2/3/4. The concrete platfo
 
 After the watch command, route by structured `readiness.state`:
 
-- `ready`: required checks passed and the same head is explicitly mergeable → SKILL step 7.
-- `checks-failed`: a required check failed or was cancelled → CI healing in SKILL step 3.
+- `ready`: all checks passed and the same head is explicitly mergeable → SKILL step 7.
+- `checks-failed`: a check failed or was cancelled → CI healing in SKILL step 3.
 - `conflicting`: the platform explicitly reports a head/base conflict → rebase healing in SKILL step 3.
 - `pending|timed-out|cancelled`: no reliable success fact exists → SKILL step 4.
 

package/templates/.agents/skills/watch-pr/reference/monitor-and-heal.zh-CN.md

@@ -6,8 +6,8 @@
 
 按 `.agents/rules/pr-checks-commands.md` 的监控命令执行后，依其退出码归为三类：
 
-- `ready`：同一 head 的 required checks 通过且平台明确可合入 → SKILL 步骤 7。
-- `checks-failed`：至少一个 required check 失败或取消 → SKILL 步骤 3 的 CI 自愈。
+- `ready`：同一 head 的全部 checks 通过且平台明确可合入 → SKILL 步骤 7。
+- `checks-failed`：至少一个 check 失败或取消 → SKILL 步骤 3 的 CI 自愈。
 - `conflicting`：平台明确当前 head 与 base 冲突 → SKILL 步骤 3 的 rebase 自愈。
 - `pending|timed-out|cancelled`：mergeability/checks 未形成可靠成功事实 → SKILL 步骤 4。
 

package/templates/.agents/templates/task.en.md

@@ -80,7 +80,7 @@ pr_status: pending             # PR status: pending (default) | created (PR crea
 
 ## Human Rulings
 
-<!-- Use ai decide <task-ref> <ordinal|ledger-id> [--needs-implementation true|false] <decision> for needs-human-decision rulings; code-stage rulings must explicitly declare whether implementation is required. -->
+<!-- Use ai decide <task-ref> <ordinal|ledger-id> [--needs-implementation true|false] <decision> for needs-human-decision rulings; upstream AI predeclares implementation intent for new code decisions, while the explicit option remains for legacy tasks or consistency checks. -->
 
 ## Implementation Inputs
 

package/templates/.agents/templates/task.zh-CN.md

@@ -80,7 +80,7 @@ pr_status: pending             # PR 状态：pending（默认）| created（已
 
 ## 人工裁决
 
-<!-- 使用 ai decide <task-ref> <序号|账本ID> [--needs-implementation true|false] <裁决内容> 写入 needs-human-decision 裁定；code 阶段必须显式声明是否需要实现。 -->
+<!-- 使用 ai decide <task-ref> <序号|账本ID> [--needs-implementation true|false] <裁决内容> 写入 needs-human-decision 裁定；新建 code 裁决的实现意图由上游 AI 预先声明，显式参数仅用于历史任务兼容或一致性复核。 -->
 
 ## 实现输入
 

package/templates/.claude/commands/watch-pr.en.md

@@ -1,5 +1,5 @@
 ---
-description: "Watch a PR's required checks and self-heal on failure"
+description: "Watch all checks on a PR and self-heal on failure"
 usage: "/watch-pr [task-ref | --task <ref> | --pr <number>]"
 ---
 

package/templates/.claude/commands/watch-pr.zh-CN.md

@@ -1,5 +1,5 @@
 ---
-description: "监控 PR 的 required checks 并在失败时自愈"
+description: "监控 PR 的全部 checks 并在失败时自愈"
 usage: "/watch-pr [task-ref | --task <ref> | --pr <number>]"
 ---
 

package/templates/.claude/rules/tool-preferences.en.md

@@ -0,0 +1,20 @@
+# Claude Code - Tool Preferences
+
+## Preferred tools
+
+| Operation | Prefer | Avoid |
+|-----------|--------|-------|
+| Find files | `Glob` | `find`, `ls` |
+| Search content | `Grep` | `grep`, `rg` |
+| Read files | `Read` | `cat`, `head`, `tail` |
+| Edit files | `Edit` | `sed`, `awk` |
+| Create files | `Write` | `echo >`, `cat <<EOF` |
+
+Use **Bash only** for Git operations, builds/tests, and system information.
+
+## Slash Commands
+
+Commands are discovered automatically from `.claude/commands/`. Type `/` at the prompt to view the complete list and descriptions.
+
+A typical task workflow is:
+`/create-task` -> `/analyze-task` -> `/plan-task` -> `/code-task` -> `/review-code` -> `/complete-task`

package/templates/.claude/rules/tool-preferences.zh-CN.md

@@ -0,0 +1,20 @@
+# Claude Code - 工具偏好
+
+## 工具使用偏好
+
+| 操作 | 推荐 | 不推荐 |
+|------|------|--------|
+| 文件搜索 | `Glob` | `find`、`ls` |
+| 内容搜索 | `Grep` | `grep`、`rg` |
+| 读取文件 | `Read` | `cat`、`head`、`tail` |
+| 编辑文件 | `Edit` | `sed`、`awk` |
+| 创建文件 | `Write` | `echo >`、`cat <<EOF` |
+
+**Bash 仅用于**：Git 操作、构建/测试、系统信息查询
+
+## Slash Commands
+
+可用命令从 `.claude/commands/` 自动发现。在提示符中输入 `/` 查看完整列表和描述。
+
+任务工作流的典型顺序：
+`/create-task` -> `/analyze-task` -> `/plan-task` -> `/code-task` -> `/review-code` -> `/complete-task`

package/templates/.opencode/commands/watch-pr.en.md

@@ -1,5 +1,5 @@
 ---
-description: "Watch a PR's required checks and self-heal on failure"
+description: "Watch all checks on a PR and self-heal on failure"
 agent: general
 subtask: false
 ---

package/templates/.opencode/commands/watch-pr.zh-CN.md

@@ -1,5 +1,5 @@
 ---
-description: "监控 PR 的 required checks 并在失败时自愈"
+description: "监控 PR 的全部 checks 并在失败时自愈"
 agent: general
 subtask: false
 ---

package/templates/.gemini/commands/_project_/review-pr.en.toml

@@ -1,10 +0,0 @@
-description = "Review a pull request by evidence grade and publish a formal review"
-prompt = """
-Review pull request {{args}}.
-
-The PR number may be given with `--pr <number>`, via a PR URL, or by reverse lookup from the current branch when omitted.
-
-Read and execute the review-pr skill from `.agents/skills/review-pr/SKILL.md`.
-
-Follow all steps defined in the skill exactly.
-"""

package/templates/.gemini/commands/_project_/review-pr.zh-CN.toml

@@ -1,10 +0,0 @@
-description = "按证据分级审查 PR 并发布正式 Review"
-prompt = """
-审查 PR {{args}}。
-
-PR 编号可通过 `--pr <number>` 指定，或使用 PR URL；省略时按当前分支反查。
-
-读取并执行 `.agents/skills/review-pr/SKILL.md` 中的 review-pr 技能。
-
-严格按照技能中定义的所有步骤执行。
-"""