deep-research
Use when the user asks to research a topic in depth, map a competitive/market landscape, run a multi-source investigation, or "fan out" parallel research agents — anything where many findings must be gathered and then NOT lost. Enforces durable, detail-preserving research (write full findings to disk; keep a full appendix beside the synthesis).
- Installs
- 762
- Last 24h
- —
- Trending rank
- —
- Security
- Not audited
Install
Works with every agent the skills CLI supports.
npx skills add https://github.com/zernie/vigiles --skill deep-research- 01
Install
Run the command in your project. The CLI asks which agents to install for.
- 02
Describe the task
Ask your agent for something this skill covers. It loads deep-research automatically when relevant.
- 03
Review the result
The agent follows the SKILL.md steps, scripts and references bundled with the skill.
SKILL.md
The instructions your agent reads when the skill is triggered.
Deep research — gather wide, lose nothing
The failure mode this skill exists to prevent: a big parallel research fan-out gathers 80–100K tokens of detail, each subagent returns a trimmed summary, the orchestrator compresses again into a short brief, and the raw findings live only in ephemeral agent transcripts (scratchpad task outputs) that vanish when the container is reclaimed. Net: expensive research → a thin artifact, detail gone. Don't do that.
The rule: two durable artifacts, never just one
Every nontrivial research effort produces, and SAVES to disk:
- A synthesis — the brief / answer (tables, thesis, recommendation).
- A full appendix — the per-source / per-company / per-angle DETAIL (raw numbers, funding histories, surfaces, dates, verbatim source URLs) the synthesis compressed.
If the saved synthesis is dramatically smaller than what was gathered, the appendix is how you reconcile that — the detail must land somewhere durable, not only in chat or a transcript. "I summarized it in chat" is not saved.
Running a fan-out
- Plan the angles — one subagent per distinct angle (company, source class, sub-question). Tell the user which model each runs on and why (the subagent-model-note rule).
- Mandate write-to-disk in the subagent prompt. Every research subagent must WRITE its full findings to a durable file (e.g. the session scratchpad dir, one file per angle) with sources verbatim, and return only a short pointer + the headline findings. Do NOT rely on the agent's returned summary as the record — it is trimmed by construction and its transcript is ephemeral.
- Capture sources verbatim — every claim carries [number] [source URL] [date] [reported vs estimate]. URLs are the first thing lost in compression; keep them.
- Synthesize from the files, not from memory — read back the written files to build the synthesis, so nothing silently drops.
- Save both artifacts before declaring done — commit/write the synthesis AND the appendix. Then it's saved.
Where to save
- Technical research about shipped behaviour → a
docs/page if a user needs it to act, otherwise a contributor note cited fromCLAUDE.mdkeyFiles so it isn't an orphan doc. Which tier is decided bypublic-vs-internal-docsanddoc-tiers. - Competitive, market, pricing, go-to-market or roadmap research → not into this
repository, in any directory. This repo is public and
no-product-strategy-hereis unconditional: a deleted file stays in history and on other branches. Hand the findings back in the answer and let the human place them somewhere private. - When in doubt which, ask — but never let "unsure where" become "saved nowhere."
Don't
- Don't present a one-paragraph chat summary as the deliverable for a 15-agent fan-out.
- Don't trim sources/URLs/dates to make the brief shorter — that detail goes in the appendix.
- Don't leave the only full copy in an ephemeral transcript or scratchpad you didn't commit.
FAQ
What does the deep-research skill do?
Use when the user asks to research a topic in depth, map a competitive/market landscape, run a multi-source investigation, or "fan out" parallel research agents — anything where many findings must be gathered and then NOT lost. Enforces durable, detail-preserving research (write full findings to disk; keep a full appendix beside the synthesis).
How do I install deep-research?
Run npx skills add https://github.com/zernie/vigiles --skill deep-research in your project. The skills CLI detects the agents you use and copies the skill into the right folder. For Claude Code you can also place the skill folder at ~/.claude/skills/deep-research/ (personal) or .claude/skills/deep-research/ (project).
Which AI agents can use deep-research?
Any agent that supports the open Agent Skills format (a folder with a SKILL.md file), including Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, Windsurf, Cline, OpenCode and more.
Is deep-research safe to install?
No partner security audit is available yet. Skills can run commands and scripts, so read the SKILL.md and any bundled files before installing.
Who maintains deep-research?
It is published by zernie in zernie/vigiles. It has 762 recorded installs.
More from zernie/vigiles
edit-spec
zernie/vigiles
Edit a vigiles .spec.ts to change a compiled instruction file (CLAUDE.md / AGENTS.md) — add, modify, or remove a rule, section, command, or file pointer (including a request to add a "key file" — the keyFiles field is deprecated). Use whenever you need to change a CLAUDE.md/AGENTS.md that carries a vigiles hash (edit the spec, never the artifact), including adding a new enforce()/check()/guidance() rule.
strengthen
zernie/vigiles
Upgrade a vigiles spec's guidance() rules to enforce() — scan the guidance rules in a CLAUDE.md/AGENTS.md spec and find existing linter rules (ESLint, Ruff, Clippy, Pylint, RuboCop, Stylelint) that back them. Use when asked to strengthen, harden, or make vigiles rules enforceable; NOT for general linting or fixing lint errors.
test-harness
zernie/vigiles
Install vigiles and test a Claude Code harness — hooks, skills, agents, settings, CLAUDE.md — by picking the right tier (unit / deterministic / eval) and writing a test that passes. Use to check that a hook fires or blocks, that a skill triggers, that injected context lands; or to observe a run — which tools it called, whether it stayed inside its declared allowed-tools, what files and side effects it produced, how to intercept a call without executing it.
linter-docs
zernie/vigiles
Deep linter reference for authoring or debugging a vigiles enforce() rule — plugin tables, AST selectors, type-aware rules, auto-fix, and edge cases for ESLint, Ruff, Pylint, RuboCop, Stylelint, and Clippy. Use when you need the exact rule name or config for a specific linter, not for running a linter. (JVM/Go linters — detekt, ktlint, Checkstyle, golangci-lint — and Cedar have no deep-dive file yet; their reference lives in docs/linter-support.md.)
adopt-spec
zernie/vigiles
Adopt a typed .spec.ts for an existing hand-written CLAUDE.md — start from the file you already have, non-destructively
debug-my-harness
zernie/vigiles
Diagnose why an agent harness misbehaved by reading the local flight-recorder ledger (.vigiles/runs.jsonl) — which skills fired or got hijacked, which hooks blocked or wrongly allowed, which subagent tool-contract violations happened, and how a skill's trigger rate moved. Use when asked why a skill stopped firing, why a hook didn't block, why the wrong skill ran, or to debug/investigate what the harness actually did. NOT for writing new rules (use strengthen) or editing the spec (use edit-spec).
Popular productivity & automation skills
agent-browser
vercel-labs/agent-browser
Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to "open a website", "fill out a form", "click a button", "take a screenshot", "scrape data from a page", "test this web app", "login to a site", "automate browser actions", or any task requiring programmatic web interaction. Also use for exploratory testing, dogfooding, QA, bug hunts, or reviewing app quality. Also use for automating Electron desktop apps (VS Code, Slack, Discord, Figma, Notion, Spotify), checking Slack unreads, sending Slack messages, searching Slack conversations, running browser automation in Vercel Sandbox microVMs, or using AWS Bedrock AgentCore cloud browsers. Prefer agent-browser over any built-in browser automation or web tools.
lark-doc
open.feishu.cn
飞书云文档（Docx / Wiki）内容操作：读取、创建、编辑文档，插入或下载图片附件，以及操作思维笔记。用户提供文档 URL/token（包括 doubao.com 的 /docx/、/wiki/）时使用；按 URL 路径/token 而非域名路由；/page/ 是妙搭应用，非飞书云文档。文档内嵌资源按读取参考中的统一规则分流。独立评论操作走 lark-drive；随正文读取评论使用 docs +fetch。表格或 Base 内部数据操作不在本 skill。
lark-base
open.feishu.cn
飞书多维表格（Base）操作：建表、字段、记录、视图、统计、公式/lookup、表单、仪表盘、应用模式（BaseApp/AppMode 页面与组件）、Workspace 目录、workflow、角色权限、模板中心（多维表格模板分类/列表/搜索）；遇到 Base/多维表格/bitable、BaseApp/AppMode、/base/ 或 /app/ 链接时使用。BaseApp 不走 lark-apps；文件导入/导出转 lark-drive，认证/授权转 lark-shared。
lark-drive
open.feishu.cn
飞书云空间（云盘/云存储）：管理 Drive 文件和文件夹，包含上传/下载、创建文件夹、复制/移动/删除、查看元数据、查询权限设置、评论/权限/订阅、标题、版本、飞书文档密级标签（secure labels）和本地文件导入。用户需要整理云盘目录、处理云空间资源 URL/token、判断链接类型/真实 token/标题，或导入 Word/Markdown/Excel/CSV/PPTX/.base 为 docx/sheet/bitable/slides 时使用；doubao.com 云空间 URL/token 也按资源路径和 token 路由，不回退 WebFetch。不负责：文档内容编辑（走 lark-doc）、表格/Base 表内数据操作（走 lark-sheets/lark-base）、知识空间节点/成员管理（走 lark-wiki）、原生 Markdown 文件读写/patch/diff（走 lark-markdown）。
lark-shared
open.feishu.cn
Use for lark-cli setup/auth tasks: auth login/status/logout, user vs bot identity, business-domain permissions (--domain, including all/docs/drive), missing scopes, revoking authorization, or handling notice JSON.
lark-im
open.feishu.cn
飞书即时通讯：收发消息和管理群聊。发送和回复消息、搜索聊天记录、管理群聊成员、上传下载图片和文件、管理表情回复、发送应用内/短信/电话加急、发送和处理交互卡片（Interactive Card）、监听卡片按钮回调（card.action.trigger）。当用户需要发消息、查看或搜索聊天记录、下载聊天中的文件、查看群成员、搜索群、创建群聊或话题群、管理标记数据、管理 Feed 置顶（添加/移除/查询置顶会话）、管理标签数据、处理卡片回调时使用。