Inspectrum

Set up Inspectrum's Codex plan gate. Use normal approvals only — do not
switch to Bypass Permissions or Full Access.

Steps:
1. Run `node --version`. If < 20, stop and tell me to install Node 20+
   from https://nodejs.org first.
2. Run `codex --version`. If it is missing or older than 0.99.0, run
   `npm install -g @openai/codex@latest` and verify again.
3. Run `claude plugin marketplace add yannmenec/inspectrum`, then
   `claude plugin install inspectrum@inspectrum`.
4. Run `claude mcp add --transport stdio --scope user inspectrum -- npx -y inspectrum@latest`
   (this powers the on-demand /inspectrum:review command).
5. Run `npx -y inspectrum@latest doctor` and show me the output,
   including the resolved codex model/effort lines.
6. Run `codex login status`.
   - If it prints "Logged in", tell me: "✅ Setup complete. Your next
     plan-mode plan gets a Codex review automatically." Done.
   - Otherwise (not logged in), pop open a Terminal window with codex
     already running by executing:
       `osascript -e 'tell application "Terminal" to do script "codex"'`
     Then tell me: "A Terminal window opened with codex running.
     Inside that window, click 'Sign in with ChatGPT', complete the
     login in my browser, then close the Terminal window. Inspectrum
     is then ready."

Do NOT use sudo, edit shell profiles, push git changes, read .env or
credentials, or publish packages. Report back: Node version, codex
install status, doctor verdict, and whether login was needed.

将以下内容粘贴到 Codex app 或 CLI 中：

Set up Inspectrum so I can review my plans with Claude. Use normal
approvals only.

Steps:
1. Run `node --version`. If < 20, stop and tell me to install Node 20+
   from https://nodejs.org first.
2. Run `claude --version`. If "command not found", run
   `npm install -g @anthropic-ai/claude-code` and verify again.
3. If `~/.inspectrum/config.toml` does not exist, create it with:

   [defaults]
   reviewers = ["claude"]

4. Run `codex mcp add inspectrum -- npx -y inspectrum@latest`
5. Run `npx -y inspectrum@latest doctor` and show me the output.
6. Pop open a Terminal window with claude already running so I can
   confirm or complete login, by executing:
     `osascript -e 'tell application "Terminal" to do script "claude"'`
   Then tell me:
   - "If claude shows its chat prompt, you're already logged in —
     close the Terminal window. Inspectrum is ready."
   - "If claude shows /login or opens a browser, complete the sign-in
     with your Claude account, then close the Terminal window.
     Inspectrum is then ready."
   (The ⚠ claude line in the doctor stays even after login because
   claude doesn't expose a status command we can detect — harmless.)

Do NOT use sudo, edit shell profiles, push git changes, read .env or
credentials, or publish packages. Report back: Node version, claude
install status, doctor verdict, and whether login was needed.

需要Claude Pro/Max 订阅，无需 API 密钥。

[defaults]
reviewers = ["codex", "gemini"]   # called in parallel
judge     = "codex"               # consolidates when >= 2 reviewers
focus     = "all"                 # correctness | completeness | risk | clarity | all

[plan_gate]                       # ExitPlanMode hook behavior
enabled          = true
max_rounds       = 2              # denials before the plan passes through
reason_max_chars = 3000           # budget for findings fed back to Claude
# reviewers      = ["codex"]      # gate-specific override of defaults.reviewers

[reviewers.codex]
effort          = "high"          # passed as -c model_reasoning_effort=high
timeout_seconds = 300             # per-reviewer override of limits.timeout_seconds
# model         = "gpt-5.6-sol"   # passed as -m; omit to inherit ~/.codex/config.toml

[reviewers.gemini]
type   = "cli"
binary = "gemini"
model  = "gemini-2.5-pro"

[reviewers.muse]
type   = "cli"
binary = "muse"                   # muse exec, prompt passed via --prompt-file
# model  = "..."                  # omit to inherit muse's own default
# effort = "high"                 # passed as --reasoning-effort

[reviewers.opencode]
type   = "cli"
binary = "opencode"               # opencode run, prompt on stdin
# model  = "google/gemini-2.5-flash"   # -m, provider/model form
# effort = "high"                 # passed as --variant

[reviewers.agy]
type   = "cli"
binary = "agy"                    # schema-constrained structured output
# model  = "..."                  # passed as --model
# effort = "high"                 # --effort; only some models accept it

[reviewers.grok]
type   = "cli"
binary = "grok"                   # confined to a read-only tool allow-list
# model  = "grok-4.6-build"       # passed as -m
# effort = "high"                 # passed as --effort

[reviewers.local]
type     = "http"
backend  = "ollama"
endpoint = "http://localhost:11434"
model    = "qwen2.5:0.5b"

[limits]
report_max_chars = 8000           # caps the stored report
timeout_seconds  = 300            # default reviewer wallclock

没有配置文件时，将使用 reviewers = ["codex"]。无需 API 密钥的设置：codex CLI 可通过 ChatGPT 登录使用，而 agy 后端是 Google 的 Antigravity CLI（curl -fsSL https://antigravity.google/cli/install.sh | bash），使用个人 Google 账号登录——这是已停用的 gemini CLI 之后可用的免费方案。实验性后端：kimi、qwen、openrouter、ollama（本地运行，不产生出站流量）。muse 后端运行 muse exec，并通过 --disable-shell --disable-write 限制权限（待审阅的计划属于不可信输入）；args 中不接受 --yolo、--disable-sandbox 和 --enable-shell-tool。opencode 后端运行 opencode run --agent summary，该权限集会禁止使用工具（默认的 build Agent 会直接运行计划文本中的 shell 命令）；args 中不接受 --agent、--auto 和 --format。agy 后端（Antigravity CLI）通过 --json-schema 限制输出，因此会在源头检查审阅结果是否符合架构；只有在你设置了 --effort 时才会发送它（部分模型不接受此参数）。agy 是 inspectrum 无法限制权限的唯一后端：经验证，--sandbox 和 --mode plan 均无法阻止写入（agy 1.1.27），因此只应将 agy 用于你信任的计划——muse、opencode 和 grok 是受限权限的选项；每当配置了 agy 时，inspectrum doctor 都会发出警告。grok 后端通过 --tools 允许列表（read_file,grep,list_dir）限制权限：经验证，--sandbox 配置和 --disallowed-tools 均无法阻止 shell 写入，而允许列表可以。args 中不接受 --tools、--always-approve 和 --permission-mode。gemini 后端已弃用，且将被移除——请迁移到 agy。

运行无头模式的主机或 CI 环境无法进行交互式登录？请改为通过 MCP 主机的 env 块传入对应的 API 密钥——OPENAI_API_KEY（codex）、ANTHROPIC_API_KEY（claude）、GEMINI_API_KEY（gemini）。手动编写的 JSON/TOML 示例位于 examples/ 下。

# Register the MCP server without the plugin:
claude mcp add --transport stdio --scope user inspectrum -- npx -y inspectrum@latest   # Claude Code
codex mcp add inspectrum -- npx -y inspectrum@latest                                   # Codex

# The automatic plan gate must use the plugin's pinned fail-open shim; do not
# register plan-gate through a mutable npm tag.

# Verify everything:
npx -y inspectrum@latest doctor

• 会话日志保存在 ~/.inspectrum/sessions/<timestamp>__<id>/，其中包含完整计划，以及每位审阅者的判定和发现的 Markdown 记录。在 POSIX 系统上，目录权限为 0700。早于 0.1.0 版本写入的日志会保留原有权限——请使用 chmod -R 700 ~/.inspectrum/sessions/ 进行修正。
• 切勿将机密粘贴到计划或上下文中。 计划会写入本地会话日志，且计划和上下文都会发送给每位活跃的审阅者。
• 云端路由：claude → Anthropic（OAuth 密钥链或 ANTHROPIC_API_KEY）；codex → OpenAI（ChatGPT 登录或 OPENAI_API_KEY）；gemini → Google（使用个人账号登录 CLI，或设置 GEMINI_API_KEY）；openrouter → openrouter.ai；kimi → Moonshot AI（设备验证码登录：kimi login，而非环境变量）；qwen → Alibaba Cloud；muse → CLI 自身存储的凭据（muse login）；opencode → 你配置的任意提供商（opencode providers login）；agy → Google Antigravity（首次运行时通过浏览器登录，或在 ~/.gemini/antigravity-cli/settings.json 中设置 modelProvider: "gemini" 并使用 GEMINI_API_KEY）；grok → xAI（Grok Build 余额）；ollama → 仅限 localhost，不产生出站流量，除非你更改 endpoint。
• Codex 会在一次性临时目录中通过 codex exec --ephemeral --skip-git-repo-check -s read-only … 调用——沙箱固定为只读，会剥除配置中的弱化沙箱和覆盖 cwd 的参数，并且 codex 不会持久化任何会话文件。