deep-research

deep-research is a skill for Claude Code, Codex from xerrors/Yuxi. It costs 69 tokens per session (1,367 once invoked), scanned A, original, MIT.

A method for organising research that uses several sources and needs fact checking. It breaks a broad question into smaller investigations, checks important claims and combines the results into a cited report.

In plain words
What is it for?
Use it for literature reviews, industry or competitor research, technology comparisons and other reports that require multiple sources, verification and citations.
Why use it?
It reduces scattered research, unsupported claims and missed contradictions. It also makes the evidence and gaps behind a conclusion easier to trace.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one.

Good fit Use it for literature reviews, industry or competitor research, technology comparisons and other reports that require multiple sources, verification and citations.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/xerrors/yuxi/deep-research
About the project

Yuxi is a self-hosted, multi-tenant knowledge-agent platform that combines document retrieval, knowledge graphs, multi-agent workflows, tools, sandboxes, and access controls in one workspace. Teams use it to build source-backed question answering and multi-step agents while managing their data, models, and permissions. The catalogue entries provide agents, skills, and instructions for Yuxi workflows.

xerrors/Yuxi · 7,405 stars · on GitHub · xerrors.github.io

Install

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.

Any agent
npx skills add xerrors/Yuxi --skill deep-research
Clone the repo
git clone --depth 1 https://github.com/xerrors/Yuxi

Made for: Claude Code, Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for deep-research

README.md
[![agentmods](https://agentmods.dev/badge/skills/xerrors/yuxi/deep-research/github.svg)](https://agentmods.dev/skills/xerrors/yuxi/deep-research)
Your own site
<a href="https://agentmods.dev/skills/xerrors/yuxi/deep-research"><img src="https://agentmods.dev/badge/skills/xerrors/yuxi/deep-research/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for deep-research

Your own site · 80×15
<a href="https://agentmods.dev/skills/xerrors/yuxi/deep-research"><img src="https://agentmods.dev/badge/skills/xerrors/yuxi/deep-research.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 69 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,367 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe. Third-party audits
  • NVIDIA SkillSpector pass 7 Sept 2026
How audits are shown
Origin original No closer match found in the catalogue.
Token cost

What 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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00069 $0.01367
Opus 5.5 $0.00028 $0.00547
Sonnet 5.5 $0.00014 $0.00273
Haiku 4.5 $0.00007 $0.00137

Measured 27d ago against content hash 0b301cba0c1b, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

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 27d 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.

backend/package/yuxi/agents/skills/buildin/deep-research/SKILL.md · 61 lines

How it starts

The opening of the file, as written. The whole thing — 61 lines — stays where its author put it; the contents beside it link to each section on GitHub.

深度研究技能

当任务目标是产出多来源、可追溯、经过核验的深度研究结论(科研综述、行业/竞品调研、技术选型、专题分析等)时,使用此技能组织整个研究过程。本技能的核心是编排:你负责整体把控与子智能体调度,把繁重的检索与核验工作派发出去,自己专注规划与综合。

可用子智能体

通过 subagent_start 工具派发,记录返回的 run_id;需要结果时调用 subagent_await。互不依赖的子任务先全部派发,再等待结果:

  • research-explorer(调研探索员):围绕一个明确子问题做多轮网页/知识库检索,返回按要点组织、带 <cite> 引用的结构化发现。这是主力,按子问题并行多开。
  • fact-verifier(事实核查员):对给定的关键论断做对抗式核验,逐条给出 支持 / 存疑 / 反驳 + 依据来源 + 置信度,并标注冲突。

编排流程

1. 澄清范围

问题不明确时,先用 ask_user_question 补充 2-3 个关键问题(研究目标、受众、范围边界、地域/时效、输出语言与形式),对齐验收标准后再开工。已经清晰的任务不要反复追问。

2. 规划拆解

用 write_todos 把研究目标拆成可独立调研的子问题,每个子问题写明产出标准(要回答什么、需要哪类证据)。子问题应正交、覆盖完整,避免重叠或遗漏关键角度。

3. 并行派发调研

  • 把互不依赖的子问题用多个 subagent_start 调用并行派发给 research-explorer。
  • 每次派发在 description 中写清:子问题目标、已知上下文、期望输出格式(要点 + <cite source="$URL" type="url">$INDEX</cite> 引用 + 参考来源列表)。
  • 何时派发 vs 自己直检:子问题复杂、需多轮检索、可隔离上下文、可并行时一律派发子智能体;仅在澄清范围、补一两个零散事实、或快速校正方向时才自己少量直接检索。
  • 子问题之间有依赖时,先派发前置子问题,拿到结果后再派发后续。

4. 核验关键结论

对影响最终结论的关键论断、数字、以及子智能体之间相互冲突的发现,派发 fact-verifier 做对抗式核验。要求其默认倾向「证据不足即标注存疑」。核验未通过的结论不要写进正文,或必须明确降级标注。

5. 综合成稿

证据充分后,由你统一综合为结构化报告,不要简单拼接子智能体返回的原文。根据研究主题组织内容,围绕论证而不是资料堆砌。

6. 停止准则

信息饱和、或确认无法获取更多有效信息即停。明确标注证据缺口与不确定性,不臆断、不编造来源。

引用

  • 子智能体返回中间证据时沿用 <cite source="$URL" type="url">$INDEX</cite>;最终报告中的关键结论、数据和观点用指向原始来源的普通 <a href="..."> 链接就近标注。
  • 引用用户附件或知识库时标明文件名或路径。

输出

  • 最终交付的是一份可直接使用的报告,而不是「我打算怎么研究」。
  • 不要外泄中间推理过程、原始检索日志,也不要把待办清单原样输出成正文。
  • 报告语言与用户提问语言一致,使用正式、克制、可复核的书面表达。
  • 用户没有指定格式时,默认把完整报告写入当前 Workdir 的 outputs/ 目录,保存为一个 .html 文档;不要把完整报告直接铺在聊天正文里。用户明确要求 Markdown、纯文本、JSON 或其他格式时,以用户要求为准。
  • 从系统提示读取当前 Project Workdir 的绝对虚拟路径,用 write_file 写入 <当前 Workdir 绝对路径>/outputs/<文件名>.html。不要把 outputs/report.html、宿主机路径或 /outputs/report.html 传给文件工具。
  • 写入成功后,把 write_file 返回的同一绝对路径传给 present_artifacts。最终聊天回复只需简要说明报告已完成并提示交付物,不要重复报告正文。

默认 HTML 应简洁、优雅并专注阅读,同时给模型保留版式判断空间:

  • 页面与正文宽度随视口自适应。宽屏可以使用侧栏目录帮助浏览章节,窄屏隐藏或折叠侧栏,让正文自然占满可用宽度。
  • 可以按内容需要引用外部样式、字体、图片或其他公开资源;图片应服务于研究内容,而不是纯装饰。
  • 不提供固定 HTML 模板;根据研究主题选择合适的章节、图表、表格和视觉层级,整体保持克制、清晰,不做复杂浮夸的展示页。

Read the full file on GitHub · 61 lines

Changes

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.

  1. 27d ago Changed · +7 lines 0b301cba0c1b
  2. 28d ago Changed · +3 lines 72b97ba83115
  3. 1mo ago First seen · 51 lines · 69 tokens per session scan A 72842a1054e0

Subscribe to this mod's changes

deep-research is a skill published in the GitHub repository xerrors/Yuxi (7,405 stars, last pushed yesterday), licensed MIT. It adds 69 tokens to every session and 1,367 once invoked, about $0.0003 per session on Opus 5.5. 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.

Related

Other skills, from other repositories

scope-creep-detector

Analyzes git diffs against a stated intent to detect scope creep, unrelated files, broad pull requests, changes that grew beyond a fix, dependency additions, public API renames, config or CI edits, oversized hunks, and formatting-only files. Use when the user asks whether a change grew beyond the fix, a PR is too…

Shubhamsaboo/awesome-llm-apps · 98 tokens

reviewing-blogs

Reviews and critiques blog posts in src/content/blog/posts/ against Appwrite's established standards for voice, tone, structure, SEO, formatting, and accuracy. Use when the user asks to review, audit, critique, or improve an existing blog post.

appwrite/appwrite · 55 tokens

first-reader

Beta readers for any draft, run by simulating how a real reader experiences it, moment by moment. A skim gate, a no-lookahead timed read producing an attention transcript with quit points, a recall test of what a reader remembers the next day, and a trust ledger of the implied author. Use when the user asks for a beta…

Shubhamsaboo/awesome-llm-apps · 174 tokens

mcp-apps-builder

MANDATORY for ALL MCP server work - mcp-use framework best practices and patterns. READ THIS FIRST before any MCP server work, including: Creating new MCP servers Modifying existing MCP servers (adding/updating tools, resources, prompts, widgets) Debugging MCP server issues or errors Reviewing MCP server code for…

Shubhamsaboo/awesome-llm-apps · 139 tokens

project-graveyard

Scans the developer's machine for dead side projects, autopsies each one from its git history (died at the payments wall, killed by a newer project, finished but never shipped), surfaces their personal death patterns, and picks the corpse most worth resurrecting — then helps ship it. Use when the user mentions…

Shubhamsaboo/awesome-llm-apps · 127 tokens

status

Show whether Mem0 memory is working in this repository, covering configuration, capture state, pending flushes, and whether the Mem0 API key is valid. Use when the user asks whether memory is on, why a memory is missing, or anything looks broken.

mem0ai/mem0 · 54 tokens