codex-deep-search深度搜索

Deep web search using Codex CLI for complex queries that need multi-source synthesis. Use when web_search (Brave) returns insufficient results, when the user asks for in-depth research, comprehensive analysis, or says "deep search", "详细搜索", "帮我查一下", or when a topic needs following multiple links and cross-referencing sources.

作者: admin | 来源: ClawHub

Codex Deep Search

Use Codex CLI's web search capability for research tasks needing more depth than Brave API snippets.

When to Prefer Over web_search

- Complex/niche topics needing multi-source synthesis
User explicitly asks for thorough/deep research
Brave results are too shallow or missing context

Usage

Dispatch Mode (recommended — background + callback)

CODEBLOCK0

After dispatch: tell user search is running, results will arrive via Telegram. Do NOT poll.

Synchronous Mode (short queries only)

CODEBLOCK1

Then read the output file and summarize.

Parameters

Flag	Required	Default	Description
INLINECODE0	Yes	—	Research query
INLINECODE1

Result Files

File	Content
INLINECODE10	Search report (incremental)
INLINECODE11

Task metadata + status | | data/codex-search-results/task-output.txt | Raw Codex output |

Key Design

- Incremental writes — results saved after each search round, survives OOM/timeout
Low reasoning effort — reduces memory, prevents OOM SIGKILL
Timeout protection — auto-stops runaway searches
Dispatch pattern — background execution with Telegram callback, no polling

Codex Deep Search

利用Codex CLI的网络搜索能力，完成需要比Brave API摘要更深入的研究任务。

何时优先使用（相较于web_search）

- 需要多源综合的复杂/小众主题
用户明确要求进行彻底/深入研究
Brave结果过于浅显或缺乏上下文

使用方法

调度模式（推荐——后台+回调）

bash
nohup bash /home/ubuntu/clawd/skills/codex-deep-search/scripts/search.sh \
--prompt 你的研究查询 \
--task-name notebooklm-research \
--telegram-group -5006066016 \
--timeout 120 > /tmp/codex-search.log 2>&1 &

调度后：告知用户搜索正在运行，结果将通过Telegram送达。请勿轮询。

同步模式（仅限短查询）

bash
bash /home/ubuntu/clawd/skills/codex-deep-search/scripts/search.sh \
--prompt 快速事实查询 \
--output /tmp/search-result.md \
--timeout 60

然后读取输出文件并进行总结。

参数

标志	必需	默认值	描述
--prompt	是	—	研究查询
--output

结果文件

文件	内容
data/codex-search-results/<任务名>.md	搜索报告（增量式）
data/codex-search-results/latest-meta.json

任务元数据+状态 | | data/codex-search-results/task-output.txt | Codex原始输出 |

关键设计

- 增量写入 — 每轮搜索后保存结果，可应对内存不足/超时
低推理开销 — 减少内存占用，防止因内存不足被SIGKILL终止
超时保护 — 自动停止失控搜索
调度模式 — 后台执行+Telegram回调，无需轮询

codex-deep-search深度搜索