repo-harness 0.9.0 → 0.9.2

package/README.es.md

@@ -85,7 +85,7 @@ artifacts.
 ## Novedades
 
 Las notas de versión viven en [`docs/CHANGELOG.md`](docs/CHANGELOG.md). La línea
-actual es `0.9.0`.
+actual es `0.9.2`.
 
 ## Cómo funciona
 
@@ -220,7 +220,7 @@ bun add -g repo-harness
 repo-harness install
 
 # Fallback con npx, con Bun ya en PATH porque el CLI corre sobre Bun
-npx -y repo-harness install
+npx -y repo-harness@latest install
 ```
 
 </details>
@@ -419,8 +419,8 @@ Guards habituales:
 
 ## Release actual
 
-- npm package: `repo-harness@0.9.0`
-- Generated workflow stamp: `repo-harness@0.9.0+template@0.9.0`
+- npm package: `repo-harness@0.9.2`
+- Generated workflow stamp: `repo-harness@0.9.2+template@0.9.2`
 - GitHub repository: `Ancienttwo/repo-harness`
 - Release history: [`docs/CHANGELOG.md`](docs/CHANGELOG.md)
 

package/README.fr.md

@@ -85,7 +85,7 @@ l'emportent.
 ## Nouveautés
 
 Les notes de version vivent dans [`docs/CHANGELOG.md`](docs/CHANGELOG.md). La
-ligne actuelle est `0.9.0`.
+ligne actuelle est `0.9.2`.
 
 ## Comment ça marche
 
@@ -224,7 +224,7 @@ bun add -g repo-harness
 repo-harness install
 
 # Fallback npx, avec Bun déjà sur PATH car le CLI s'exécute sur Bun
-npx -y repo-harness install
+npx -y repo-harness@latest install
 ```
 
 </details>
@@ -424,8 +424,8 @@ Guards courants :
 
 ## Release actuelle
 
-- npm package : `repo-harness@0.9.0`
-- Generated workflow stamp : `repo-harness@0.9.0+template@0.9.0`
+- npm package : `repo-harness@0.9.2`
+- Generated workflow stamp : `repo-harness@0.9.2+template@0.9.2`
 - GitHub repository : `Ancienttwo/repo-harness`
 - Release history : [`docs/CHANGELOG.md`](docs/CHANGELOG.md)
 

package/README.ja.md

@@ -75,7 +75,7 @@ review、checks、handoff と食い違う場合は、source artifacts を優先
 ## What's New
 
 リリースノートは [`docs/CHANGELOG.md`](docs/CHANGELOG.md) にあります。現在の
-ラインは `0.9.0` です。
+ラインは `0.9.2` です。
 
 ## 仕組み
 
@@ -201,7 +201,7 @@ bun add -g repo-harness
 repo-harness install
 
 # npx fallback。CLI runtime は Bun なので、Bun が PATH 上に必要です。
-npx -y repo-harness install
+npx -y repo-harness@latest install
 ```
 
 </details>
@@ -398,8 +398,8 @@ hook がブロックしたときは、まず terminal の構造化された出
 
 ## 現在の Release
 
-- npm package：`repo-harness@0.9.0`
-- Generated workflow stamp：`repo-harness@0.9.0+template@0.9.0`
+- npm package：`repo-harness@0.9.2`
+- Generated workflow stamp：`repo-harness@0.9.2+template@0.9.2`
 - GitHub repository：`Ancienttwo/repo-harness`
 - Release history：[`docs/CHANGELOG.md`](docs/CHANGELOG.md)
 

package/README.md

@@ -85,7 +85,7 @@ active plan, contract, review, checks, or handoff, the source artifacts win.
 ## What's New
 
 Release notes live in [`docs/CHANGELOG.md`](docs/CHANGELOG.md). The current line
-is `0.9.0`.
+is `0.9.2`.
 
 ## How It Works
 
@@ -234,6 +234,10 @@ concrete sprint instead of reinterpreting the original chat.
 
 ## First 5 Minutes
 
+<p align="center">
+  <img src="docs/images/repo-harness-install-donkey-carrot.png" alt="Pixel art donkey following a carrot for repo-harness installation" width="900">
+</p>
+
 This is the fastest path for an AI tooling owner evaluating whether the workflow is
 safe to adopt in a real repo. It separates the machine-level runtime bootstrap
 from the repo-local contract install, so a dry run can show exactly what will
@@ -256,20 +260,20 @@ curl -fsSL https://raw.githubusercontent.com/Ancienttwo/repo-harness/main/instal
 irm https://raw.githubusercontent.com/Ancienttwo/repo-harness/main/install.ps1 | iex
 ```
 
-<details>
-<summary>Already have Bun? Use Bun directly, or npx as a fallback</summary>
+If Bun is already on PATH, you can skip the shell installer:
 
 ```bash
-# Bun (recommended)
+# Bun one-shot bootstrap
+bunx repo-harness@latest install
+
+# Or install the persistent CLI first
 bun add -g repo-harness
 repo-harness install
 
 # npx fallback, with Bun already on PATH because the CLI runs on Bun
-npx -y repo-harness install
+npx -y repo-harness@latest install
 ```
 
-</details>
-
 ### 2. Bootstrap the host runtime once
 
 ```bash
@@ -279,18 +283,24 @@ repo-harness install
 `install` is the first-run global bootstrap path. It installs the current npm
 package as the global CLI, refreshes repo-harness skill aliases, installs
 user-level hook adapters, configures Waza runtime skills, persists a brain root
-under `~/.repo-harness/config.json`, and configures CodeGraph MCP. In an
-interactive terminal it asks Y/n before installing the external skills and
-CodeGraph pieces (Enter keeps today's default of installing both); non-TTY
-runs and `--json` stay unprompted with the same default-on behavior. It does
-not apply repo-local workflow files to the current directory. `repo-harness
-init` remains a compatibility alias for existing scripts.
+under `~/.repo-harness/config.json`, and configures CodeGraph MCP. The command
+is idempotent: when the CLI is already installed from Bun's global package
+source, it skips the CLI reinstall and still refreshes the host runtime pieces.
+In an interactive terminal it asks Y/n before installing the external skills
+and CodeGraph pieces (Enter keeps today's default of installing both); non-TTY
+runs and `--json` stay unprompted with the same default-on behavior. Passing
+`--no-external-skills` or `--no-codegraph` explicitly also skips that item's
+prompt unprompted, which is the escape hatch for PTY-allocating CI (for example
+`docker run -t`). It does not apply repo-local workflow files to the current
+directory. `repo-harness init` remains a compatibility alias for existing
+scripts.
 
 For an Agent-owned, read-only bootstrap audit, run `repo-harness setup check
---json` or add `--check-updates` for version advisories. `setup check` is
-not a runtime hook: it does not write user-level files, install updates, or
-register adapters. It emits `agent_actions` with the reason, risk, target files,
-optional command, and verification surface for the Agent to execute deliberately.
+--json` or add `--check-updates` for version and adopted-repo refresh
+advisories. `setup check` is not a runtime hook: it does not write user-level
+files, install updates, run `adopt`, or register adapters. It emits
+`agent_actions` with the reason, risk, target files, optional command, and
+verification surface for the Agent to execute deliberately.
 `repo-harness init-hook` remains a compatibility alias.
 
 ### Install and refresh examples
@@ -312,6 +322,17 @@ repo-harness update --check
 repo-harness adopt --repo /path/to/repo
 ```
 
+`repo-harness install --target codex|both --location global` also resolves and
+persists the Codex delegation mode read by the delegation-advisor hook: pass
+`--delegation-mode auto|explicit` for non-interactive use, or answer the
+one-line interactive prompt shown in a TTY (Enter keeps the current choice,
+defaulting to `explicit` when nothing is configured yet). The chosen mode is
+written to `delegation.mode` in `~/.repo-harness/config.json`, merged with
+existing keys such as `brainRoot`; this global value takes precedence over a
+repo's `.ai/harness/policy.json` `delegation.mode`. The step never fires for
+`--target claude`, and with no flag and no TTY it leaves the file untouched —
+it never writes a silent default.
+
 ### 3. Preview the repo-local contract
 
 ```bash
@@ -612,8 +633,8 @@ Most common guards:
 
 ## Current Release
 
-- npm package: `repo-harness@0.9.0`
-- Generated workflow stamp: `repo-harness@0.9.0+template@0.9.0`
+- npm package: `repo-harness@0.9.2`
+- Generated workflow stamp: `repo-harness@0.9.2+template@0.9.2`
 - GitHub repository: `Ancienttwo/repo-harness`
 - Release history: [`docs/CHANGELOG.md`](docs/CHANGELOG.md)
 

package/README.zh-CN.md

@@ -75,7 +75,7 @@ review、checks 或 handoff 冲突，以 source artifacts 为准。
 
 ## What's New
 
-Release notes 见 [`docs/CHANGELOG.md`](docs/CHANGELOG.md)，当前版本线是 `0.9.0`。
+Release notes 见 [`docs/CHANGELOG.md`](docs/CHANGELOG.md)，当前版本线是 `0.9.2`。
 
 ## 工作原理
 
@@ -198,6 +198,10 @@ source of truth，Codex Goal mode 只围绕具体 sprint 恢复和推进，而
 
 ## 前 5 分钟
 
+<p align="center">
+  <img src="docs/images/repo-harness-install-donkey-carrot.png" alt="repo-harness 安装引导的 pixel art 驴和萝卜 banner" width="900">
+</p>
+
 这是评估一个真实仓库是否适合接入该 workflow 的最快路径。它把机器级 runtime
 bootstrap 和 repo-local contract install 分开，所以 dry-run 能先展示会改什么，
 再决定是否应用。
@@ -218,20 +222,20 @@ curl -fsSL https://raw.githubusercontent.com/Ancienttwo/repo-harness/main/instal
 irm https://raw.githubusercontent.com/Ancienttwo/repo-harness/main/install.ps1 | iex
 ```
 
-<details>
-<summary>已经有 Bun？优先直接用 Bun，也可以把 npx 作为备选</summary>
+如果 Bun 已经在 `PATH` 上，可以跳过 shell installer：
 
 ```bash
-# Bun（推荐）
+# Bun 一步 bootstrap
+bunx repo-harness@latest install
+
+# 或者先安装持久化 CLI
 bun add -g repo-harness
 repo-harness install
 
 # npx 备选；仍要求 Bun 已在 PATH 上，因为 CLI runtime 是 Bun
-npx -y repo-harness install
+npx -y repo-harness@latest install
 ```
 
-</details>
-
 ### 2. 先做一次 host runtime bootstrap
 
 ```bash
@@ -240,13 +244,16 @@ repo-harness install
 
 `install` 是首次全局引导入口。它把当前 npm 包安装成全局 CLI，刷新 repo-harness
 skill aliases，安装 user-level hook adapters，配置 Waza runtime skills，把 brain
-root 持久化到 `~/.repo-harness/config.json`，并配置 CodeGraph MCP。它不会把当前目录
-默认迁移成 repo-local workflow。`repo-harness init` 保留为兼容 alias，给已有脚本用。
+root 持久化到 `~/.repo-harness/config.json`，并配置 CodeGraph MCP。这个命令是
+幂等的：如果 CLI 已经来自 Bun global package source，它会跳过 CLI 重装，但仍继续
+刷新 host runtime pieces。它不会把当前目录默认迁移成 repo-local workflow。
+`repo-harness init` 保留为兼容 alias，给已有脚本用。
 
 如果要让 Agent 做只读 bootstrap audit，运行 `repo-harness setup check
---json`；需要版本提示时加 `--check-updates`。`setup check` 不是 runtime hook：
-它不会写 user-level files、安装更新或注册 adapters，只输出带 reason、risk、
-targets、可选 command 和 verification 的 `agent_actions`，由 Agent 再显式执行。
+--json`；需要版本和已接入仓库刷新提示时加 `--check-updates`。`setup check`
+不是 runtime hook：它不会写 user-level files、安装更新、执行 `adopt` 或注册
+adapters，只输出带 reason、risk、targets、可选 command 和 verification 的
+`agent_actions`，由 Agent 再显式执行。
 `repo-harness init-hook` 保留为兼容 alias。
 
 ### 安装和刷新例子
@@ -451,8 +458,8 @@ hook block 工作时，先看 terminal 里的结构化输出。核心字段是
 
 ## 当前 Release
 
-- npm package：`repo-harness@0.9.0`
-- Generated workflow stamp：`repo-harness@0.9.0+template@0.9.0`
+- npm package：`repo-harness@0.9.2`
+- Generated workflow stamp：`repo-harness@0.9.2+template@0.9.2`
 - GitHub repository：`Ancienttwo/repo-harness`
 - Release history：[`docs/CHANGELOG.md`](docs/CHANGELOG.md)
 

package/assets/AGENTS.md

@@ -37,7 +37,7 @@ Owns the workflow-engine-contract-assets capability boundary declared in .ai/con
 - Architecture domain: `workflow-engine`
 - Architecture capability: `contract-assets`
 - Architecture module: `docs/architecture/modules/workflow-engine/contract-assets.md`
-- Last architecture event: 2026-07-05T04:35:33+0800
+- Last architecture event: 2026-07-06T13:54:27+0800
 - Last changed path: `.ai/harness/policy.json`
 - Severity: high
 - Change type: workflow-surface

package/assets/CLAUDE.md

@@ -37,7 +37,7 @@ Owns the workflow-engine-contract-assets capability boundary declared in .ai/con
 - Architecture domain: `workflow-engine`
 - Architecture capability: `contract-assets`
 - Architecture module: `docs/architecture/modules/workflow-engine/contract-assets.md`
-- Last architecture event: 2026-07-05T04:35:33+0800
+- Last architecture event: 2026-07-06T13:54:27+0800
 - Last changed path: `.ai/harness/policy.json`
 - Severity: high
 - Change type: workflow-surface

package/assets/hooks/AGENTS.md

@@ -49,8 +49,8 @@ Owns the runtime-harness-hook-adapters capability boundary declared in .ai/conte
 - Architecture domain: `runtime-harness`
 - Architecture capability: `hook-adapters`
 - Architecture module: `docs/architecture/modules/runtime-harness/hook-adapters.md`
-- Last architecture event: 2026-07-05T13:45:11+0800
-- Last changed path: `assets/hooks/codex-delegation-advisor.sh`
+- Last architecture event: 2026-07-09T11:39:06+0800
+- Last changed path: `assets/hooks/prompt-guard.sh`
 - Severity: high
 - Change type: workflow-surface
 - Module responsibility: Keep this block aligned with the local boundary described by surrounding human-owned context.

package/assets/hooks/CLAUDE.md

@@ -49,8 +49,8 @@ Owns the runtime-harness-hook-adapters capability boundary declared in .ai/conte
 - Architecture domain: `runtime-harness`
 - Architecture capability: `hook-adapters`
 - Architecture module: `docs/architecture/modules/runtime-harness/hook-adapters.md`
-- Last architecture event: 2026-07-05T13:45:11+0800
-- Last changed path: `assets/hooks/codex-delegation-advisor.sh`
+- Last architecture event: 2026-07-09T11:39:06+0800
+- Last changed path: `assets/hooks/prompt-guard.sh`
 - Severity: high
 - Change type: workflow-surface
 - Module responsibility: Keep this block aligned with the local boundary described by surrounding human-owned context.

package/assets/hooks/codex-delegation-advisor.sh

@@ -56,6 +56,19 @@ JSON_INPUT="$input" REPO_ROOT="${HOOK_REPO_ROOT:-$(pwd)}" bun -e '
     return null;
   }
 
+  function readGlobalDelegationMode() {
+    try {
+      const home = process.env.HOME;
+      if (!home) return null;
+      const raw = fs.readFileSync(path.join(home, ".repo-harness", "config.json"), "utf8");
+      const parsed = JSON.parse(raw);
+      const mode = parsed && parsed.delegation && parsed.delegation.mode;
+      return mode === "auto" || mode === "explicit" ? mode : null;
+    } catch {
+      return null;
+    }
+  }
+
   let input;
   try {
     input = JSON.parse(process.env.JSON_INPUT || "");
@@ -91,7 +104,6 @@ JSON_INPUT="$input" REPO_ROOT="${HOOK_REPO_ROOT:-$(pwd)}" bun -e '
     { name: "chinese-subagent", pattern: /交给\s*子代理|使用多个\s*(agent|代理)|并行(调查|研究|处理|执行|agent|代理)/i },
   ];
   const trigger = triggers.find((entry) => entry.pattern.test(prompt) && !(entry.skipDiscussion && isDelegationDiscussion(prompt)));
-  if (!trigger) process.exit(0);
 
   const repoRoot = process.env.REPO_ROOT || process.cwd();
 
@@ -102,6 +114,16 @@ JSON_INPUT="$input" REPO_ROOT="${HOOK_REPO_ROOT:-$(pwd)}" bun -e '
   } catch {
     policyDelegation = {};
   }
+
+  const globalDelegationMode = readGlobalDelegationMode();
+  const effectiveDelegationMode = globalDelegationMode || (policyDelegation.mode === "auto" ? "auto" : "explicit");
+
+  const explicit = Boolean(trigger);
+  if (!explicit) {
+    if (effectiveDelegationMode !== "auto") process.exit(0);
+    if (isDelegationDiscussion(prompt)) process.exit(0);
+  }
+
   const maxAgents = Number.isInteger(policyDelegation.max_agents) ? policyDelegation.max_agents : 3;
   const maxDepth = Number.isInteger(policyDelegation.max_depth) ? policyDelegation.max_depth : 1;
   const preferredRunners =
@@ -122,17 +144,17 @@ JSON_INPUT="$input" REPO_ROOT="${HOOK_REPO_ROOT:-$(pwd)}" bun -e '
   const state = {
     version: 1,
     eligible: true,
-    explicit: true,
+    explicit,
     spawned: false,
     fallback_used: false,
-    mode: "explicit",
+    mode: explicit ? "explicit" : "auto",
     max_agents: maxAgents,
     max_depth: maxDepth,
     allow_parallel_writers: false,
-    stop_fallback: true,
+    stop_fallback: explicit,
     preferred_runners: preferredRunners,
     fallback_runner: fallbackRunner,
-    trigger: trigger.name,
+    trigger: explicit ? trigger.name : "auto-mode",
     prompt_hash: crypto.createHash("sha1").update(prompt).digest("hex"),
     scope_source: scope?.source || "unscoped",
     scope_id: scope?.id || "",
@@ -151,7 +173,9 @@ JSON_INPUT="$input" REPO_ROOT="${HOOK_REPO_ROOT:-$(pwd)}" bun -e '
   const context = [
     "[repo-harness:delegation]",
     "",
-    "The current user prompt explicitly enabled bounded delegation.",
+    explicit
+      ? "The current user prompt explicitly enabled bounded delegation."
+      : "Repo policy delegation.mode=auto is standing user authorization for bounded delegation, selected by the user at install time. This satisfies the user-authorization requirement for native subagents (spawn_agent) when the dispatch conditions below are met.",
     "",
     "Treat the active task contract (tasks/contracts/<active-plan-stem>.contract.md) as the authoritative execution brief: Goal, Scope, Allowed Paths, and Exit Criteria. Do not re-derive scope from this conversation.",
     "",

package/assets/hooks/prompt-guard.sh

@@ -1054,7 +1054,7 @@ prompt_guard_engine_call
 if [[ "$PG_ENGINE_STATE" != "ok" ]]; then
   if [[ "$PG_ENGINE_STATE" == "legacy" ]]; then
     echo "[PromptGuard] Advisory: the installed repo-harness CLI predates the prompt-verdict protocol; prompt intent gates are degraded to advisory for this prompt."
-    echo "[PromptGuard] Refresh the CLI with: bun add -g repo-harness@latest && repo-harness install. Fallback: npx -y repo-harness install."
+    echo "[PromptGuard] Refresh the CLI with: bun add -g repo-harness@latest && repo-harness install. Fallback: npx -y repo-harness@latest install."
   else
     echo "[PromptGuard] Advisory: prompt-guard decision engine is unavailable (repo-harness CLI or bun not found); prompt intent gates are degraded to advisory for this prompt."
     echo "[PromptGuard] Edit-layer guards still enforce plan and contract scope. Install the repo-harness CLI to restore prompt decisions."

package/assets/reference-configs/external-tooling.md

@@ -41,7 +41,9 @@ brain config, package-manager globals, or user-authored sibling hook entries.
 runtime by default. Third-party tooling and CodeGraph registration stay
 readiness findings from `repo-harness setup check` unless the update command is
 run with an explicit opt-in such as `--with-external-skills` or
-`--configure-codegraph`.
+`--configure-codegraph`. Repo-local workflow refresh stays on
+`repo-harness adopt`; `setup check --check-updates` reports an Agent action when
+the current adopted repo's dry-run adoption plan has pending operations.
 
 The cross-review skills are **harness-owned and self-contained** — their source
 lives in `assets/skills/<skill>/` and they wrap the peer CLI (`codex exec` /
@@ -121,13 +123,13 @@ cd ~/.claude/skills/gstack && ./setup --host codex
 Both hosts:
 
 ```bash
-npx -y skills add tw93/Waza -g -a claude-code codex -s think hunt check health -y
+bunx skills add tw93/Waza -g -a claude-code codex -s think hunt check health -y
 ```
 
 Single host:
 
 ```bash
-npx -y skills add tw93/Waza -g -a claude-code -s think hunt check health -y
+bunx skills add tw93/Waza -g -a claude-code -s think hunt check health -y
 ```
 
 Replace `claude-code` with `codex` when installing for Codex only.
@@ -156,13 +158,13 @@ done
 Both hosts:
 
 ```bash
-npx -y skills add hylarucoder/hai-stack -g -a claude-code codex -s geju -y
+bunx skills add hylarucoder/hai-stack -g -a claude-code codex -s geju -y
 ```
 
 Single host:
 
 ```bash
-npx -y skills add hylarucoder/hai-stack -g -a claude-code -s geju -y
+bunx skills add hylarucoder/hai-stack -g -a claude-code -s geju -y
 ```
 
 Replace `claude-code` with `codex` when installing for Codex only.
@@ -204,7 +206,9 @@ command directly unless local policy explicitly opts into vendoring.
 ### Runtime Ownership Boundary
 
 `repo-harness setup check --target <host> --check-updates --json` reports the
-execution base as separate `runtime.*` checks. Keep the boundary explicit:
+execution base as separate `runtime.*` checks and reports repo-local adoption
+refresh as `repo.adopt-refresh` when the current repo has opted in. Keep the
+boundary explicit:
 
 | Capability | Owner | Required for |
 |---|---|---|
@@ -424,7 +428,7 @@ cd ~/.claude/skills/gstack && git pull && ./setup --host codex
 ### Waza
 
 ```bash
-npx -y skills update
+bunx skills update
 for d in think hunt check health; do
   rsync -a --delete ~/.agents/skills/$d/ ~/.codex/skills/$d/
 done
@@ -443,7 +447,7 @@ done
 ### hai-stack
 
 ```bash
-npx -y skills update
+bunx skills update
 rsync -a --delete ~/.agents/skills/geju/ ~/.codex/skills/geju/
 diff -qr ~/.agents/skills/geju ~/.codex/skills/geju
 ```
@@ -491,10 +495,15 @@ guessed mapping.
 
 | Upstream frontmatter | Codex TOML |
 |---|---|
-| `model: opus`, `effort: max` | `model = "gpt-5.5"`, `model_reasoning_effort = "xhigh"` |
-| `model: sonnet`, `effort: max` | `model = "gpt-5.5"`, `model_reasoning_effort = "medium"` |
+| `model: opus`, `effort: max` | `model = "gpt-5.6-sol"`, `model_reasoning_effort = "xhigh"` |
+| `model: sonnet`, `effort: max` | `model = "gpt-5.6-terra"`, `model_reasoning_effort = "medium"` |
 | `tools: [...]` present | `sandbox_mode = "read-only"` |
 
+The Codex generator also rewrites the exact upstream provider label in the
+description (`Opus 4.8 at max effort` or `Sonnet 5 at max effort`) to the
+mapped GPT-5.6 model and reasoning level. A missing label fails closed so the
+installed metadata cannot claim a different model from the TOML settings.
+
 `developer_instructions` is the upstream `.md` body plus the canonical
 EXECUTION_BOUNDARY anti-extras clause, kept byte-identical to the
 `EXECUTION_BOUNDARY` constant in `scripts/contract-run.ts` so every generated

package/assets/reference-configs/global-working-rules.md

@@ -5,6 +5,12 @@ Use this content for user-level `~/.codex/AGENTS.md` and `~/.claude/CLAUDE.md` w
 ```md
 # Global Working Rules
 
+Rule 0: You may spend as much time as needed thinking. Do not send optional commentary progress messages. Use tools only when they are required. For tasks that do not require tools, complete the reasoning first, then answer in final.
+
+Reasoning: Prefer first principles over pattern matching. Before solving, first identify the observable and controllable conditions. For quantitative logic problems, before the final answer, you must prove the strategy is sufficient in the worst case. Numeric answers must have their arithmetic rechecked.
+
+Generality: These are general working rules. Do not tailor behavior to any specific evaluation or expected answer.
+
 - Use Chinese by default for this user; keep technical terms in English. If the user writes in another language, mirror that language.
 - Act as an engineering collaborator: finish the concrete task, verify it, then report conclusion, actual change, reason, verification, and residual risk.
 - Prefer direct execution over repeated confirmation. Stop to ask only when continuing would likely produce output contrary to the user's intent.

package/assets/skill-version.json

@@ -1,6 +1,6 @@
 {
-  "version": "0.9.0",
-  "templateVersion": "0.9.0",
+  "version": "0.9.2",
+  "templateVersion": "0.9.2",
   "skillName": "repo-harness",
   "contractId": "tasks-first-harness-v1",
   "compatibility": {
@@ -191,6 +191,14 @@
     {
       "version": "0.9.0",
       "description": "Adds a bugfix task profile with root-cause evidence gating, distills contract-brief prompts and templates, makes the agent fleet and geju first-class dependencies, adds a frontend task profile and archcontext boundary bridge, and lands the EXECUTION_BOUNDARY anti-extras clause across delegated runners"
+    },
+    {
+      "version": "0.9.1",
+      "description": "Hardens the Bun-first installer and setup-check path with bunx skills probes, PATH readback hints, and explicit Codex delegation-mode selection"
+    },
+    {
+      "version": "0.9.2",
+      "description": "Hardens CLI environment diagnostics and hook execution intent while aligning the generated GPT-5.6 agent fleet with current Codex runtime guidance"
     }
   ],
   "generatedProjectStamp": {

package/assets/templates/helpers/architecture-event.ts

@@ -645,6 +645,13 @@ function findLatestMatchingFile(dir: string, token: string, extension: string):
   return files.at(-1) || "(none yet)";
 }
 
+function activePendingRequestFile(file: string): string {
+  if (!file || file === "unknown") return "(none)";
+  if (file.includes("/archive/")) return "(none)";
+  if (!existsSync(file)) return "(none)";
+  return isPendingRequest(file) ? file : "(none)";
+}
+
 function collectFiles(dir: string): string[] {
   if (!existsSync(dir)) return [];
   const files: string[] = [];
@@ -680,7 +687,7 @@ function renderContractBlock(args: Args): string {
   const severity = args.options.severity || "unknown";
   const changeType = args.options.changeType || "unknown";
   const lspProfile = args.options.lspProfile || "typescript-lsp";
-  const requestFile = args.options.requestFile || "unknown";
+  const requestFile = activePendingRequestFile(args.options.requestFile || "");
 
   return [
     "<!-- BEGIN ARCHITECTURE CONTRACT -->",

package/assets/templates/helpers/capability-resolver.ts

@@ -27,9 +27,9 @@ export type CapabilityRegistry = {
 
 // archcontext-boundaries-v1 is a deliberately narrow, read-only export of the
 // capability registry shaped as a subset of ArchitectureNode (archcontext.node/v1).
-// See tests/fixtures/archcontext/architecture-node.subset.schema.json for the
-// vendored schema subset and docs/researches/20260705-archcontext-capability-filing-handover.md
-// for why this stays a read-only bridge (no archctx runtime dependency).
+// Tests derive the bridge subset from archctx-contracts' authoritative schema;
+// docs/researches/20260705-archcontext-capability-filing-handover.md explains
+// why this stays a read-only bridge with no archctx CLI/daemon runtime dependency.
 export type ArchContextBoundaryNode = {
   schemaVersion: "archcontext.node/v1";
   id: string;

package/assets/templates/helpers/check-agent-tooling.sh

@@ -709,7 +709,7 @@ function inspectWazaSkill(host, skill, skillLock, skillItems, upstreamSkills) {
 function detectWaza() {
   const skillLockPath = path.join(HOME, ".agents", ".skill-lock.json");
   const skillLock = readJson(skillLockPath);
-  const skillsResult = run("npx", ["-y", "skills", "ls", "-g", "--json"], { timeoutMs: 1500 });
+  const skillsResult = run("bunx", ["skills", "ls", "-g", "--json"], { timeoutMs: 1500 });
   const skillItems = skillsResult.ok ? parseJson(skillsResult.stdout) || [] : [];
   const wazaEntries = Object.entries(skillLock?.skills || {}).filter(([, meta]) => meta?.source === WAZA_SOURCE_REPO);
   const upstream = fetchWazaUpstreamSkills();
@@ -815,7 +815,7 @@ function detectWaza() {
         : "Local Waza SKILL.md and shared rule files match upstream GitHub raw content.";
 
   const status = summarizeWazaStatus(hostStatuses);
-  const installCommand = `npx -y skills add tw93/Waza -g -a ${
+  const installCommand = `bunx skills add tw93/Waza -g -a ${
     hostMode === "both" ? "claude-code codex" : hostMode === "claude" ? "claude-code" : "codex"
   } -s think hunt check health -y`;
   const syncCommand = SELECTED_HOSTS.map((host) => buildWazaHostSyncCommand(host)).join(" && ");
@@ -878,8 +878,8 @@ function detectRuntimeCapabilities(waza) {
     ),
     npx: commandCapability(
       "npx",
-      "external Skills CLI bootstrap/update commands for Waza and Mermaid",
-      "external-skills-cli",
+      "no repo-harness usage; Skills CLI bootstrap/update for Waza and Mermaid runs through bunx instead",
+      "npm-registry",
       false
     ),
     skills_cli: {
@@ -889,7 +889,7 @@ function detectRuntimeCapabilities(waza) {
       owner: "external-skills-cli",
       required: false,
       required_for: "Waza/Mermaid external skill bootstrap; repo-harness reports this as an explicit exception boundary",
-      command: "npx -y skills ls -g --json",
+      command: "bunx skills ls -g --json",
     },
     bash: commandCapability(
       "bash",

package/assets/templates/helpers/context-contract-sync.sh

@@ -155,6 +155,19 @@ metadata_value() {
   ' "$file" 2>/dev/null
 }
 
+active_pending_request_file() {
+  local file="$1"
+  if [[ -z "$file" || "$file" == "unknown" || "$file" == *"/archive/"* || ! -f "$file" ]]; then
+    echo "(none)"
+    return 0
+  fi
+  if [[ "$(metadata_value "$file" "Status" | tr '[:upper:]' '[:lower:]')" == "pending" ]]; then
+    echo "$file"
+  else
+    echo "(none)"
+  fi
+}
+
 format_active_workstreams() {
   local dir="$1"
   local count=0
@@ -440,6 +453,7 @@ if [[ "$latest_snapshot" != "(none yet)" ]]; then
   semantic_diagram_source="$latest_snapshot"
 fi
 active_workstreams="$(format_active_workstreams "$workstream_dir")"
+pending_request_file="$(active_pending_request_file "${request_file:-}")"
 
 block_tmp="$(mktemp)"
 cat > "$block_tmp" <<EOF_BLOCK
@@ -466,7 +480,7 @@ cat > "$block_tmp" <<EOF_BLOCK
 - Latest snapshot: \`${latest_snapshot}\`
 - Semantic diagram source: \`${semantic_diagram_source}\`
 - Latest human diagram: \`${latest_human_diagram}\`
-- Pending architecture request: \`${request_file:-unknown}\`
+- Pending architecture request: \`${pending_request_file}\`
 
 ## Active Workstreams