菜单
Use when diagnosing Codex CLI install, config, auth, runtime, feature flags, sandbox denials, debug models, prompt input, app-server, remote-control, remote websocket/unix connections, or local Codex health failures.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Use when locating, summarizing, auditing, or debugging Codex session rollout JSONL logs by CODEX_THREAD_ID, cwd, query, issue key, project path, child or inherited-history symptoms, malformed or huge log symptoms, permission concerns, or "what happened in this Codex thread" questions.
Use when auditing agent skills or plugins for safety, coverage, duplicated logic, token/context cost, prompt-contract risk, validation readiness, marketplace visibility, dependencies, network behavior, or install/update risk. For structural split/merge/move/delete/router decisions across skills or plugins, use capability-portfolio-architect instead.
Use when auditing agent skill/plugin portfolio architecture for cross-plugin capability overlap, weak routing, duplicated or missing skills, split/merge/delete/move decisions, shared capability extraction, reference/script extraction, or plugin boundary redesign. For safety, coverage, or validation review of a single artifact without boundary changes, use capability-auditor instead.
Use when performing broad external-first discovery, vetting, scoring, distillation, or synthesis of agent skills or plugin packs from the public web, GitHub/public repositories, OpenClaw/ClawHub, marketplaces, research sources, community implementations, user references, or local skills. Use for well-vetted capability synthesis, cross-skill comparison, plugin-pack synthesis, global capability acquisition, skill strengthening, and adoption/rejection changelogs.
Route and orchestrate agent skill and plugin lifecycle work. Use when work involves agent skills, plugins, marketplace packages, capability acquisition, synthesis, creation, installation, packaging, audit, or context-density optimization.
Use when creating, updating, validating, optionally installing/cache-refreshing, or handing off marketplace-backed agent plugins. Use for plugin manifests, local marketplace entries, plugin-pack synthesis, triggerable skill bundles, install visibility gates, and Codex app deeplinks.
| name | codex-doctor-debugger |
| description | Use when diagnosing Codex CLI install, config, auth, runtime, feature flags, sandbox denials, debug models, prompt input, app-server, remote-control, remote websocket/unix connections, or local Codex health failures. |
Bundled commands use $PLUGIN_ROOT ($env:PLUGIN_ROOT in PowerShell; same path suffix) for the plugin root. Set it once: use the host's plugin-root variable when defined (Claude Code: PLUGIN_ROOT="$CLAUDE_PLUGIN_ROOT"), otherwise the absolute path of this plugin's root directory.
Use this skill when Codex CLI itself is failing or unclear: installation, config, auth, runtime health, feature flags, sandbox denials, debug model catalog, app-server transports, remote-control daemon, remote connection errors, or local environment inconsistencies.
Start with cheap, non-mutating commands:
python3 "$PLUGIN_ROOT/scripts/codex_cli_inspector.py" --commands doctor debug sandbox features app-server remote-control --json
codex doctor --summary --ascii
Use machine-readable doctor output when automation needs to classify issues:
codex doctor --json
Do not paste raw doctor output into chat if it contains local paths, config fragments, or environment details that are not needed. Summarize issue names, counts, and actionable rows.
Use codex doctor --summary --ascii first. Expand only when the summary points
to a specific area:
codex doctor --all --ascii
codex doctor --json
If config parsing is suspect, run the target command with --strict-config to
surface unsupported fields for that CLI version.
Read before writing:
codex features list
Use codex features enable <feature> or codex features disable <feature> only
when the user explicitly asks to change a flag or the issue clearly requires it.
Report any config file mutation.
Use sandbox commands to reproduce permission failures without launching a full agent session:
codex sandbox -C "$PROJECT" --log-denials -- <command> <args>
codex sandbox -C "$PROJECT" --permissions-profile <name> -- <command> <args>
On macOS, --log-denials can help identify blocked paths or sockets. Keep
allowed Unix socket paths explicit:
codex sandbox -C "$PROJECT" --allow-unix-socket ./tmp/socket -- <command>
Use debug commands for local model/config visibility:
codex debug models
codex debug prompt-input
codex debug app-server --help
Treat debug output as potentially sensitive because it can include model, config, prompt-input, or environment-derived data.
codex app-server and codex remote-control are experimental surfaces. Use
them only for app/IDE integration, remote TUI, protocol generation, or daemon
debugging.
Prefer loopback or Unix socket transports. Do not bind WebSocket listeners to non-loopback interfaces without explicit security requirements and token setup.
Examples:
codex app-server --listen unix://
codex app-server --listen ws://127.0.0.1:0 --ws-auth capability-token --ws-token-file "$TOKEN_FILE"
codex remote-control --json start
codex remote-control --json stop
Report the diagnostic lane, commands run, exact issue names or error text, version/help facts used, any config or daemon state changed, and the smallest next repair or verified healthy state.