harnessed 4.8.0 → 4.9.1

package/workflows/verify/multispec/SKILL.md

@@ -66,17 +66,18 @@ Phase-level `on` clause (critical-release 升级触发):
 
 ## How to invoke
 
-Use the Bash tool to run:
-
-```bash
-echo "$ARGUMENTS" | harnessed run verify-multispec --task-stdin
-```
-
-If `$ARGUMENTS` is empty, run `harnessed run verify-multispec` (no stdin pipe).
-
-After completion, the Bash output prints a `Next:` hint on stderr suggesting the next stage. Decide whether to invoke based on conversation context — the hint is informational, not prescriptive.
-
-<!-- harnessed-generated:v3.4.4 -->
+Orchestrated CC-natively. Do NOT pipe to `harnessed run verify-multispec` — that is the CI/headless
+path (an in-process SDK spawn that blocks the session, bypasses Agent Teams, and hangs when
+invoked from inside a Claude Code session).
+
+Run the `/verify-multispec` slash command instead (generated by `harnessed setup` at
+`~/.claude/commands/verify-multispec.md`). It drives the stage natively: `harnessed gates` → which
+subs fire, `harnessed prompt <sub>` → each spawn-ready prompt, then a CC-native subagent
+(Task / Agent tool) per fired sub, recording each outcome with `harnessed checkpoint`. The
+full state-machine steps live in `~/.claude/commands/verify-multispec.md`; if that file is absent,
+follow that same gates → prompt → spawn → checkpoint sequence yourself.
+
+<!-- harnessed-generated:v4.9.1 -->
 
 ## References
 

package/workflows/verify/multispec/SKILL.zh-Hans.md

@@ -64,19 +64,18 @@ Phase 级 `on` 子句（critical-release 升级触发）：
 - **Token 估算前置条件**：`team_cost < 2 × subagent_cost`（engine-level check；bundled cost guideline）
 - **Cleanup 必跑**：phase 02-team-cleanup `agent-teams-shutdown` 必跑（bundled cleanup discipline）
 
-## 调用方式
+## 如何调用
 
-使用 Bash 工具运行：
+CC-native 编排。**不要** pipe 到 `harnessed run verify-multispec` —— 那是 CI/headless 路径(in-process
+SDK spawn,会阻塞 session、绕过 Agent Teams,在 Claude Code 内部调用时还会挂死)。
 
-```bash
-echo "$ARGUMENTS" | harnessed run verify-multispec --task-stdin
-```
+改用 `/verify-multispec` slash command(由 `harnessed setup` 生成于 `~/.claude/commands/verify-multispec.md`)。
+它以 CC-native 方式驱动:`harnessed gates` 决定哪些 sub fire,`harnessed prompt <sub>` 给出每个
+spawn-ready prompt,然后用 CC-native subagent(Task / Agent 工具)逐个 spawn 已 fire 的 sub,
+每个结果用 `harnessed checkpoint` 记录。完整 state-machine 步骤见 `~/.claude/commands/verify-multispec.md`;
+若该文件不存在,自行按 gates → prompt → spawn → checkpoint 同序执行。
 
-若 `$ARGUMENTS` 为空，运行 `harnessed run verify-multispec`（不带 stdin pipe）。
-
-执行完成后，Bash 输出会在 stderr 打印 `Next:` 提示，建议下一阶段操作。是否继续调用，请根据对话上下文自行判断——该提示仅供参考，并非强制指令。
-
-<!-- harnessed-generated:v3.4.4 -->
+<!-- harnessed-generated:v4.9.1 -->
 
 ## 参考资料
 

package/workflows/verify/paranoid/SKILL.md

@@ -52,17 +52,18 @@ Sister `workflows/judgments/stage-routing.yaml`:
 
 ## How to invoke
 
-Use the Bash tool to run:
-
-```bash
-echo "$ARGUMENTS" | harnessed run verify-paranoid --task-stdin
-```
-
-If `$ARGUMENTS` is empty, run `harnessed run verify-paranoid` (no stdin pipe).
-
-After completion, the Bash output prints a `Next:` hint on stderr suggesting the next stage. Decide whether to invoke based on conversation context — the hint is informational, not prescriptive.
-
-<!-- harnessed-generated:v3.4.4 -->
+Orchestrated CC-natively. Do NOT pipe to `harnessed run verify-paranoid` — that is the CI/headless
+path (an in-process SDK spawn that blocks the session, bypasses Agent Teams, and hangs when
+invoked from inside a Claude Code session).
+
+Run the `/verify-paranoid` slash command instead (generated by `harnessed setup` at
+`~/.claude/commands/verify-paranoid.md`). It drives the stage natively: `harnessed gates` → which
+subs fire, `harnessed prompt <sub>` → each spawn-ready prompt, then a CC-native subagent
+(Task / Agent tool) per fired sub, recording each outcome with `harnessed checkpoint`. The
+full state-machine steps live in `~/.claude/commands/verify-paranoid.md`; if that file is absent,
+follow that same gates → prompt → spawn → checkpoint sequence yourself.
+
+<!-- harnessed-generated:v4.9.1 -->
 
 ## References
 

package/workflows/verify/paranoid/SKILL.zh-Hans.md

@@ -50,19 +50,18 @@ Sister `workflows/judgments/stage-routing.yaml`：
 - ✅ **触发**：关键模块 PR 前（auth / payment / data migration / core algorithm 等）
 - ❌ **跳过**：常规 PR / docs / config / 非核心 module
 
-## 调用方式
+## 如何调用
 
-使用 Bash 工具运行：
+CC-native 编排。**不要** pipe 到 `harnessed run verify-paranoid` —— 那是 CI/headless 路径(in-process
+SDK spawn,会阻塞 session、绕过 Agent Teams,在 Claude Code 内部调用时还会挂死)。
 
-```bash
-echo "$ARGUMENTS" | harnessed run verify-paranoid --task-stdin
-```
+改用 `/verify-paranoid` slash command(由 `harnessed setup` 生成于 `~/.claude/commands/verify-paranoid.md`)。
+它以 CC-native 方式驱动:`harnessed gates` 决定哪些 sub fire,`harnessed prompt <sub>` 给出每个
+spawn-ready prompt,然后用 CC-native subagent(Task / Agent 工具)逐个 spawn 已 fire 的 sub,
+每个结果用 `harnessed checkpoint` 记录。完整 state-machine 步骤见 `~/.claude/commands/verify-paranoid.md`;
+若该文件不存在,自行按 gates → prompt → spawn → checkpoint 同序执行。
 
-若 `$ARGUMENTS` 为空，运行 `harnessed run verify-paranoid`（不带 stdin pipe）。
-
-执行完成后，Bash 输出会在 stderr 打印 `Next:` 提示，建议下一阶段操作。是否继续调用，请根据对话上下文自行判断——该提示仅供参考，并非强制指令。
-
-<!-- harnessed-generated:v3.4.4 -->
+<!-- harnessed-generated:v4.9.1 -->
 
 ## 参考资料
 

package/workflows/verify/progress/SKILL.md

@@ -48,17 +48,18 @@ verify-progress-always trigger)。无 skip 条件 — verify-work 起点必跑
 
 ## How to invoke
 
-Use the Bash tool to run:
-
-```bash
-echo "$ARGUMENTS" | harnessed run verify-progress --task-stdin
-```
-
-If `$ARGUMENTS` is empty, run `harnessed run verify-progress` (no stdin pipe).
-
-After completion, the Bash output prints a `Next:` hint on stderr suggesting the next stage. Decide whether to invoke based on conversation context — the hint is informational, not prescriptive.
-
-<!-- harnessed-generated:v3.4.4 -->
+Orchestrated CC-natively. Do NOT pipe to `harnessed run verify-progress` — that is the CI/headless
+path (an in-process SDK spawn that blocks the session, bypasses Agent Teams, and hangs when
+invoked from inside a Claude Code session).
+
+Run the `/verify-progress` slash command instead (generated by `harnessed setup` at
+`~/.claude/commands/verify-progress.md`). It drives the stage natively: `harnessed gates` → which
+subs fire, `harnessed prompt <sub>` → each spawn-ready prompt, then a CC-native subagent
+(Task / Agent tool) per fired sub, recording each outcome with `harnessed checkpoint`. The
+full state-machine steps live in `~/.claude/commands/verify-progress.md`; if that file is absent,
+follow that same gates → prompt → spawn → checkpoint sequence yourself.
+
+<!-- harnessed-generated:v4.9.1 -->
 
 ## References
 

package/workflows/verify/progress/SKILL.zh-Hans.md

@@ -45,19 +45,18 @@ Sister `workflows/capabilities.yaml` 条目：
 总 fire 当 `phase.stage == 'verify'`（sister `workflows/judgments/stage-routing.yaml`
 verify-progress-always trigger）。无 skip 条件 — verify-work 起点必跑。
 
-## 调用方式
+## 如何调用
 
-使用 Bash 工具运行：
+CC-native 编排。**不要** pipe 到 `harnessed run verify-progress` —— 那是 CI/headless 路径(in-process
+SDK spawn,会阻塞 session、绕过 Agent Teams,在 Claude Code 内部调用时还会挂死)。
 
-```bash
-echo "$ARGUMENTS" | harnessed run verify-progress --task-stdin
-```
+改用 `/verify-progress` slash command(由 `harnessed setup` 生成于 `~/.claude/commands/verify-progress.md`)。
+它以 CC-native 方式驱动:`harnessed gates` 决定哪些 sub fire,`harnessed prompt <sub>` 给出每个
+spawn-ready prompt,然后用 CC-native subagent(Task / Agent 工具)逐个 spawn 已 fire 的 sub,
+每个结果用 `harnessed checkpoint` 记录。完整 state-machine 步骤见 `~/.claude/commands/verify-progress.md`;
+若该文件不存在,自行按 gates → prompt → spawn → checkpoint 同序执行。
 
-若 `$ARGUMENTS` 为空，运行 `harnessed run verify-progress`（不带 stdin pipe）。
-
-执行完成后，Bash 输出会在 stderr 打印 `Next:` 提示，建议下一阶段操作。是否继续调用，请根据对话上下文自行判断——该提示仅供参考，并非强制指令。
-
-<!-- harnessed-generated:v3.4.4 -->
+<!-- harnessed-generated:v4.9.1 -->
 
 ## 参考资料
 

package/workflows/verify/qa/SKILL.md

@@ -54,17 +54,18 @@ Sister `workflows/judgments/stage-routing.yaml`:
 
 ## How to invoke
 
-Use the Bash tool to run:
-
-```bash
-echo "$ARGUMENTS" | harnessed run verify-qa --task-stdin
-```
-
-If `$ARGUMENTS` is empty, run `harnessed run verify-qa` (no stdin pipe).
-
-After completion, the Bash output prints a `Next:` hint on stderr suggesting the next stage. Decide whether to invoke based on conversation context — the hint is informational, not prescriptive.
-
-<!-- harnessed-generated:v3.4.4 -->
+Orchestrated CC-natively. Do NOT pipe to `harnessed run verify-qa` — that is the CI/headless
+path (an in-process SDK spawn that blocks the session, bypasses Agent Teams, and hangs when
+invoked from inside a Claude Code session).
+
+Run the `/verify-qa` slash command instead (generated by `harnessed setup` at
+`~/.claude/commands/verify-qa.md`). It drives the stage natively: `harnessed gates` → which
+subs fire, `harnessed prompt <sub>` → each spawn-ready prompt, then a CC-native subagent
+(Task / Agent tool) per fired sub, recording each outcome with `harnessed checkpoint`. The
+full state-machine steps live in `~/.claude/commands/verify-qa.md`; if that file is absent,
+follow that same gates → prompt → spawn → checkpoint sequence yourself.
+
+<!-- harnessed-generated:v4.9.1 -->
 
 ## References
 

package/workflows/verify/qa/SKILL.zh-Hans.md

@@ -52,19 +52,18 @@ Sister `workflows/judgments/stage-routing.yaml`：
 - setup 需 Python 后端 (Tortoise ORM / pandas) → `webapp-testing` skill
 - 性能 / a11y / 内存诊断 → 不在此子工作流，用 `chrome-devtools-mcp`
 
-## 调用方式
+## 如何调用
 
-使用 Bash 工具运行：
+CC-native 编排。**不要** pipe 到 `harnessed run verify-qa` —— 那是 CI/headless 路径(in-process
+SDK spawn,会阻塞 session、绕过 Agent Teams,在 Claude Code 内部调用时还会挂死)。
 
-```bash
-echo "$ARGUMENTS" | harnessed run verify-qa --task-stdin
-```
+改用 `/verify-qa` slash command(由 `harnessed setup` 生成于 `~/.claude/commands/verify-qa.md`)。
+它以 CC-native 方式驱动:`harnessed gates` 决定哪些 sub fire,`harnessed prompt <sub>` 给出每个
+spawn-ready prompt,然后用 CC-native subagent(Task / Agent 工具)逐个 spawn 已 fire 的 sub,
+每个结果用 `harnessed checkpoint` 记录。完整 state-machine 步骤见 `~/.claude/commands/verify-qa.md`;
+若该文件不存在,自行按 gates → prompt → spawn → checkpoint 同序执行。
 
-若 `$ARGUMENTS` 为空，则运行 `harnessed run verify-qa`（不接管道输入）。
-
-完成后，Bash 输出会在 stderr 打印 `Next:` 提示，建议下一阶段。是否调用取决于对话上下文 — 该提示仅供参考，非强制指令。
-
-<!-- harnessed-generated:v3.4.4 -->
+<!-- harnessed-generated:v4.9.1 -->
 
 ## 参考资料
 

package/workflows/verify/security/SKILL.md

@@ -49,17 +49,18 @@ Sister `workflows/judgments/stage-routing.yaml`:
 
 ## How to invoke
 
-Use the Bash tool to run:
-
-```bash
-echo "$ARGUMENTS" | harnessed run verify-security --task-stdin
-```
-
-If `$ARGUMENTS` is empty, run `harnessed run verify-security` (no stdin pipe).
-
-After completion, the Bash output prints a `Next:` hint on stderr suggesting the next stage. Decide whether to invoke based on conversation context — the hint is informational, not prescriptive.
-
-<!-- harnessed-generated:v3.4.4 -->
+Orchestrated CC-natively. Do NOT pipe to `harnessed run verify-security` — that is the CI/headless
+path (an in-process SDK spawn that blocks the session, bypasses Agent Teams, and hangs when
+invoked from inside a Claude Code session).
+
+Run the `/verify-security` slash command instead (generated by `harnessed setup` at
+`~/.claude/commands/verify-security.md`). It drives the stage natively: `harnessed gates` → which
+subs fire, `harnessed prompt <sub>` → each spawn-ready prompt, then a CC-native subagent
+(Task / Agent tool) per fired sub, recording each outcome with `harnessed checkpoint`. The
+full state-machine steps live in `~/.claude/commands/verify-security.md`; if that file is absent,
+follow that same gates → prompt → spawn → checkpoint sequence yourself.
+
+<!-- harnessed-generated:v4.9.1 -->
 
 ## References
 

package/workflows/verify/security/SKILL.zh-Hans.md

@@ -47,19 +47,18 @@ Sister `workflows/judgments/stage-routing.yaml`：
 - ✅ **触发**: auth flow / session / credentials / API keys / SQL injection 路径 / OWASP top 10 area
 - ❌ **跳过**: docs / 纯 UI styling / 内部 refactor / non-security PR
 
-## 调用方式
+## 如何调用
 
-使用 Bash 工具运行：
+CC-native 编排。**不要** pipe 到 `harnessed run verify-security` —— 那是 CI/headless 路径(in-process
+SDK spawn,会阻塞 session、绕过 Agent Teams,在 Claude Code 内部调用时还会挂死)。
 
-```bash
-echo "$ARGUMENTS" | harnessed run verify-security --task-stdin
-```
+改用 `/verify-security` slash command(由 `harnessed setup` 生成于 `~/.claude/commands/verify-security.md`)。
+它以 CC-native 方式驱动:`harnessed gates` 决定哪些 sub fire,`harnessed prompt <sub>` 给出每个
+spawn-ready prompt,然后用 CC-native subagent(Task / Agent 工具)逐个 spawn 已 fire 的 sub,
+每个结果用 `harnessed checkpoint` 记录。完整 state-machine 步骤见 `~/.claude/commands/verify-security.md`;
+若该文件不存在,自行按 gates → prompt → spawn → checkpoint 同序执行。
 
-若 `$ARGUMENTS` 为空，则运行 `harnessed run verify-security`（不接管道输入）。
-
-完成后，Bash 输出会在 stderr 打印 `Next:` 提示，建议下一阶段。是否调用取决于对话上下文 — 该提示仅供参考，非强制指令。
-
-<!-- harnessed-generated:v3.4.4 -->
+<!-- harnessed-generated:v4.9.1 -->
 
 ## 参考资料
 

package/workflows/verify/simplify/SKILL.md

@@ -49,17 +49,18 @@ Sister `workflows/judgments/stage-routing.yaml`:
 
 ## How to invoke
 
-Use the Bash tool to run:
-
-```bash
-echo "$ARGUMENTS" | harnessed run verify-simplify --task-stdin
-```
-
-If `$ARGUMENTS` is empty, run `harnessed run verify-simplify` (no stdin pipe).
-
-After completion, the Bash output prints a `Next:` hint on stderr suggesting the next stage. Decide whether to invoke based on conversation context — the hint is informational, not prescriptive.
-
-<!-- harnessed-generated:v3.4.4 -->
+Orchestrated CC-natively. Do NOT pipe to `harnessed run verify-simplify` — that is the CI/headless
+path (an in-process SDK spawn that blocks the session, bypasses Agent Teams, and hangs when
+invoked from inside a Claude Code session).
+
+Run the `/verify-simplify` slash command instead (generated by `harnessed setup` at
+`~/.claude/commands/verify-simplify.md`). It drives the stage natively: `harnessed gates` → which
+subs fire, `harnessed prompt <sub>` → each spawn-ready prompt, then a CC-native subagent
+(Task / Agent tool) per fired sub, recording each outcome with `harnessed checkpoint`. The
+full state-machine steps live in `~/.claude/commands/verify-simplify.md`; if that file is absent,
+follow that same gates → prompt → spawn → checkpoint sequence yourself.
+
+<!-- harnessed-generated:v4.9.1 -->
 
 ## References
 

package/workflows/verify/simplify/SKILL.zh-Hans.md

@@ -47,19 +47,18 @@ Sister `workflows/judgments/stage-routing.yaml`：
 - ✅ **触发**: verify chain 末尾步骤 (所有其他 verify sub 已 ship，准备 code 简化收尾)
 - ❌ **跳过**: verify chain 中间步骤 (避免过早简化干扰后续 review)
 
-## 调用方式
+## 如何调用
 
-使用 Bash 工具运行：
+CC-native 编排。**不要** pipe 到 `harnessed run verify-simplify` —— 那是 CI/headless 路径(in-process
+SDK spawn,会阻塞 session、绕过 Agent Teams,在 Claude Code 内部调用时还会挂死)。
 
-```bash
-echo "$ARGUMENTS" | harnessed run verify-simplify --task-stdin
-```
+改用 `/verify-simplify` slash command(由 `harnessed setup` 生成于 `~/.claude/commands/verify-simplify.md`)。
+它以 CC-native 方式驱动:`harnessed gates` 决定哪些 sub fire,`harnessed prompt <sub>` 给出每个
+spawn-ready prompt,然后用 CC-native subagent(Task / Agent 工具)逐个 spawn 已 fire 的 sub,
+每个结果用 `harnessed checkpoint` 记录。完整 state-machine 步骤见 `~/.claude/commands/verify-simplify.md`;
+若该文件不存在,自行按 gates → prompt → spawn → checkpoint 同序执行。
 
-若 `$ARGUMENTS` 为空，则运行 `harnessed run verify-simplify`（不接管道输入）。
-
-完成后，Bash 输出会在 stderr 打印 `Next:` 提示，建议下一阶段。是否调用取决于对话上下文 — 该提示仅供参考，非强制指令。
-
-<!-- harnessed-generated:v3.4.4 -->
+<!-- harnessed-generated:v4.9.1 -->
 
 ## 参考资料