Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/feiskyer/codex-settings/deep-researchnpx skills add feiskyer/codex-settings --skill deep-researchgit clone --depth 1 https://github.com/feiskyer/codex-settingsWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00119 | $0.01285 |
| Opus 5 | $0.00060 | $0.00642 |
| Sonnet 5 | $0.00024 | $0.00257 |
| Haiku 4.5 | $0.00012 | $0.00128 |
Grade A, and why
deep-research scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 137 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Deep Research
把深度调研作为可重复、可审计的生产流程执行。主控负责目标澄清、样本摸底、子任务设计、结果核验和最终综合;子进程负责边界清晰的采集或局部分析。
核心约束
- 保持用户当前模型和推理配置,不传
--model,不覆盖无关配置。 - 子进程默认使用
workspace-write;只有确实需要 shell 网络访问时才启用 runner 的--network。 - 先检查当前会话可用的 Skills、连接器和 MCP,再按来源适配能力;不要假设固定服务或工具名存在。
- 不使用
--dangerously-bypass-approvals-and-sandbox。 - 所有运行产物写入独立的
.research/<name>/目录。 - 在开始批量执行前向用户展示拆分方案;需要明显成本、长时间运行或外部系统访问时,等待明确同意。
Bundled scripts
先解析当前 Skill 的绝对目录并记为 <skill-dir>。
scripts/run_children.py:跨平台并行执行codex exec,负责超时、重试、日志和结果状态。scripts/aggregate.py:按 manifest 顺序聚合成功的子报告,缺失或空结果时失败。
两个脚本都使用 Python 标准库,不生成临时 shell 脚本。
Workflow
1. 澄清与摸底
明确目标、受众、时间范围、来源边界、评价标准和最终格式。通过当前可用工具获取少量真实样本,记录代表性来源和缺口,避免只凭经验拆分。
2. 创建运行目录
使用不重复的语义化名称,例如:
.research/20260712-codex-skills-a3f2/
├── prompts/
├── logs/
├── child_outputs/
├── raw/
├── cache/
└── manifest.json
把网页原文、数据和解析结果缓存到 raw/ 或 cache/,避免重复抓取。
3. 设计子任务
每个子任务只负责一个明确边界,Prompt 至少包含:
- 子目标、输入和允许访问的范围
- 输出结构和证据要求
- 失败时必须说明原因,不得编造结果
- 输出自然语言 Markdown,并把来源链接放在对应结论附近
将 Prompt 分别写入 prompts/,然后创建 manifest:
{
"tasks": [
{
"id": "market-history",
"title": "市场历史",
"prompt_file": "prompts/market-history.md"
},
{
"id": "current-competitors",
"title": "当前竞品",
"prompt_file": "prompts/current-competitors.md"
}
]
}
id 只能包含字母、数字、点、下划线和短横线。Prompt 必须位于本次运行目录内。
4. 预检和执行
先预览命令,不启动 Codex 子进程:
python3 "<skill-dir>/scripts/run_children.py" \
--run-dir ".research/<name>" \
--workspace "$PWD" \
--dry-run
检查 manifest 和 Prompt 后执行:
python3 "<skill-dir>/scripts/run_children.py" \
--run-dir ".research/<name>" \
--workspace "$PWD" \
--parallel 8 \
--timeout 600 \
--retries 1
只有子任务必须通过 shell 直接联网时才添加 --network。根据任务成本调整并发和超时,先用 1–2 个子任务验证链路,再扩大并发。
Runner 固定输出:
child_outputs/<id>.mdlogs/<id>.logresults.json
5. 核验和失败处理
读取 results.json,检查失败、超时、空输出和引用缺失。只重试失败的边界任务;不要因为单个失败重新运行所有成功任务。需要改变模型、权限或来源范围时先说明原因。
What ships with it
4 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 3d ago First seen · 137 lines · 119 tokens per session scan A da27b3cf6c7f
deep-research is a skill published in the GitHub repository feiskyer/codex-settings (236 stars, last pushed 20d ago), licensed MIT. It adds 119 tokens to every session and 1,285 once invoked, about $0.0006 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
skill-creator
Create, refine, and benchmark agent skills. Use when building a new skill, updating an existing one, running evals, checking trigger quality, or improving a skill description.
deep-research
Multi-agent research orchestration: split a research goal into parallel sub-goals, run each via headless claude -p subprocesses, aggregate results into a polished report file. Use for systematic web/document research, competitive or industry analysis, batch link/dataset processing, and long-form evidence synthesis.…
github-review-pr
Review GitHub pull requests with detailed, multi-perspective code analysis using parallel subagents. Use this skill whenever the user wants to review a PR, asks for code review on a pull request, mentions "review PR", "check this PR", "look at pull request", or references a PR number or GitHub PR URL. Do NOT use for…
brainstorming
Explore user intent, requirements, and design options through collaborative dialogue before implementation. Use before building new features, components, or systems — whenever the user describes something to build and design decisions are involved. Triggers: "brainstorm", "help me design", "think through the…
grill-me
针对方案或设计的高强度追问式面试(adversarial design review / grill session),暴露假设漏洞与缺失约束,过程中同步维护领域模型(术语表和 ADR)。手动调用 /grill-me。.
gpt-image-skill
Generate or edit images using OpenAI GPT Image API (gpt-image-2, gpt-image-1, etc). Use ONLY when the user explicitly names OpenAI or GPT as the provider: "gpt image", "openai image", "generate image with openai", "用 openai 画图", "用 GPT 生成图片". For generic image requests without a provider, use nanobanana-skill instead.…