Agent skill

Auto Research

by huytieu in huytieu/COG-second-brain

Deep strategic research engine — decomposes questions into parallel research threads, spawns multiple agents, and synthesizes into actionable strategic analysis

MITAuto-check passedResearch & Science

Install Auto Research

skills CLI
$ npx skills add huytieu/COG-second-brain --skill auto-research -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install huytieu/COG-second-brain auto-research --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/huytieu/COG-second-brain.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/auto-research .claude/skills/auto-research && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
auto-research
GitHub stars
1.3k
Token cost
~2.9k tokens
SKILL.md length
948 words
Files
1
Skills in repo
37
Repo updated
First seen
Licence
MIT

At a glance

Deep strategic research engine — decomposes questions into parallel research threads, spawns multiple agents, and synthesizes into actionable strategic analysis

  • Works in 4 steps: Question Decomposition (Orchestrator — 2… → Parallel Deep Research (Spawn 5-7 Agents… → Synthesis (Orchestrator — after all… → …
  • Research & Science work in your project
  • SKILL.md covers When to Invoke, Agent Mode Awareness, Command: /auto-research and Input, plus 6 more sections
  • Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md

What it does

Auto Research is an agent skill from huytieu/COG-second-brain. Deep strategic research engine — decomposes questions into parallel research threads, spawns multiple agents, and synthesizes into actionable strategic analysis

Its SKILL.md is about 2.9k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.

It sits in Research & Science. The repository describes itself as: Self-evolving second brain with 35 AI skills, 10 agents, and people CRM. Closed-loop harness: a V-model verification lifecycle where the worker never grades its own homework… The licence is MIT.

When your agent uses it

  • Research & Science work in your project

Example prompts

  • “/auto-research”

Workflow steps

4 steps, taken from the step headings in SKILL.md.

  1. Question Decomposition (Orchestrator — 2 minutes)
  2. Parallel Deep Research (Spawn 5-7 Agents Simultaneously)
  3. Synthesis (Orchestrator — after all agents complete)
  4. Save & Deliver

What it can do on your machine

Read from SKILL.md and the folder at commit 36ac9d7. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    No scripts in the folder and no shell commands in SKILL.md (its code samples are markdown).

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Auto Research loads about 2.9k tokens when it runs. Until then it costs about 44 tokens; SKILL.md has 948 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~44
When it runs · the whole SKILL.md, loaded when a task matches
~2.9k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from huytieu/COG-second-brain at commit 36ac9d7, republished under its MIT licence (© huytieu). 948 words, ~2,920 tokens.

Download SKILL.mdSave it as .claude/skills/auto-research/SKILL.md (or your agent's skills folder).
name
auto-research
description
Deep strategic research engine — decomposes questions into parallel research threads, spawns multiple agents, and synthesizes into actionable strategic analysis
roles
product-manager, engineering-lead, founder, all

COG Auto Research Skill

When to Invoke

  • User asks a strategic question requiring deep research
  • User says "research", "auto-research", "investigate", "strategic analysis", "deep dive into [topic]"
  • User wants to understand market forces, competitive dynamics, technology trajectories, or strategic options
  • User needs evidence-based analysis with real sources to support decision-making

Inspired by Karpathy's autoresearch — but for strategic thinking instead of ML training.

Agent Mode Awareness

Check agent_mode in 00-inbox/MY-PROFILE.md frontmatter:

  • If agent_mode: team — use the full parallel agent execution strategy (5-7 agents). This skill benefits massively from team mode.
  • If agent_mode: solo — run 2-3 sequential research passes with WebSearch/WebFetch, produce a lighter analysis without the full multi-thread structure.

Command: /auto-research

Input

The user provides a strategic question or topic as the command argument. Examples:

  • "If foundation models commoditize, what happens to LLM wrapper companies like Katalon/Scout?"
  • "Future of the testing industry as AI capabilities expand"
  • "Should we build vs buy vs partner for our AI layer?"
  • "What are the strategic options for Scout if OpenAI launches a testing product?"

Execution Strategy

Phase 1: Question Decomposition (Orchestrator — 2 minutes)

Break the user's strategic question into 5-7 research threads that together will provide a comprehensive answer. Each thread should be:

  • Independent — can be researched in parallel
  • Specific — has a clear research objective
  • Complementary — together they cover the full strategic landscape

Decomposition framework:

  1. Market forces — what macro trends drive this question?
  2. Historical precedent — has this pattern played out before in other industries?
  3. Player analysis — who are the key players and what are they doing?
  4. Technology trajectory — where is the underlying tech heading?
  5. Customer behavior — what do end-users actually want/do?
  6. Economic model — what are the unit economics and value capture dynamics?
  7. Emerging tech & architectures — what concepts, projects, or frameworks are still in development/discussion (pre-mainstream) that could be foundational? Research open-source projects, research papers, GitHub repos, Discord/forum discussions, conference talks, and early-stage tools that are relevant. Examples: novel agent architectures, new testing paradigms, experimental frameworks. These may not have polished docs — dig into READMEs, GitHub issues, Twitter/X threads, blog posts from builders, and academic preprints.
  8. Contrarian view — what's the strongest argument against the consensus?

Not all threads apply to every question. Pick the 5-7 most relevant. Thread 7 (Emerging tech) should ALWAYS be included — the user specifically wants to stay ahead of concepts that aren't mainstream yet.

Before spawning agents:

  1. Read relevant files from the vault for existing context:
    • 05-knowledge/ for existing frameworks and mental models
    • 04-projects/ for project-specific context if relevant
    • Recent braindumps for the user's existing thinking on this topic
  2. State the decomposition to the user so they can course-correct before agents launch
Phase 2: Parallel Deep Research (Spawn 5-7 Agents Simultaneously)

CRITICAL: Launch ALL agents in a single message. Use run_in_background: true for all agents.

Each agent gets a detailed prompt following this template:

You are a strategic research analyst investigating a specific thread of a larger strategic question.

MAIN QUESTION: [user's original question]
YOUR THREAD: [specific research thread]
EXISTING CONTEXT: [any relevant vault context]

RESEARCH METHODOLOGY:
1. WebSearch for 8-12 high-quality sources (prioritize: research reports, expert analyses, company filings, academic papers, industry publications — NOT listicles or superficial blog posts)
2. For each source found, WebFetch to read the full content and extract key arguments, data points, and frameworks
3. Look for CONFLICTING viewpoints — don't just confirm one narrative
4. Identify specific data points, statistics, and concrete examples
5. Note the credibility and potential bias of each source
6. FOR EMERGING TECH THREADS: Go beyond polished sources. Search GitHub repos (README, issues, discussions), Twitter/X threads from builders, Discord/forum discussions, conference talk summaries, arXiv preprints, and early blog posts. The goal is to surface concepts that are pre-mainstream but technically promising. For each concept found, assess: maturity level, technical approach, relevance to the user's use case, and what it would take to adopt/integrate.

OUTPUT FORMAT (return ALL of this):

## Thread: [thread name]

### Key Findings (3-5 bullet points)
- Finding with source attribution

### Evidence & Data Points
- Specific statistics, market data, examples with sources

### Expert/Notable Perspectives
- Named perspectives from credible voices

### Implications for [user's context]
- What this means specifically for the user's situation

### Confidence Level
- HIGH / MEDIUM / LOW with reasoning

### Sources
- Numbered list of actual URLs consulted

Agent naming convention: research-[thread-slug] (e.g., research-market-forces, research-historical-precedent)

Phase 3: Synthesis (Orchestrator — after all agents complete)

Once all agents return, synthesize into a single strategic analysis document:

Document Structure:
markdown
---
type: strategic-research
domain: [auto-detect from question]
date: YYYY-MM-DD
question: "[original question]"
threads: [list of research threads]
confidence: [overall confidence HIGH/MEDIUM/LOW]
tags:
  - auto-research
  - strategy
  - [topic tags]
status: complete
---

# [Strategic Question as Title]

## Executive Summary
3-5 sentences capturing the core insight. Lead with the answer, not the process.

## The Strategic Landscape
Synthesized view across all research threads. Not a thread-by-thread dump — weave findings together into a coherent narrative.

## Key Forces at Play
The 3-4 most important dynamics shaping this question, with evidence from multiple threads.

## Scenarios
### Scenario A: [Most Likely] — X% confidence
What happens, timeline, implications

### Scenario B: [Optimistic/Alternative]
What happens, timeline, implications

### Scenario C: [Worst Case/Disruption]
What happens, timeline, implications

## Emerging Tech & Architectures to Watch
Concepts, projects, and frameworks that are still in development/discussion but could be foundational. For each:
- **What it is:** One-paragraph explanation
- **Maturity:** Pre-alpha / Alpha / Early adoption / Growing community
- **Technical approach:** How it works architecturally
- **Relevance to our use case:** Why it matters for us specifically
- **Adoption path:** What it would take to integrate/adopt — effort, risks, dependencies
- **Key links:** GitHub repo, paper, discussion thread

## Strategic Options
For each option:
- **Description:** What this means concretely
- **Pros:** With evidence
- **Cons:** With evidence
- **Prerequisites:** What needs to be true
- **Timeline:** When to decide/act
- **Emerging tech leverage:** Which emerging concepts from above could strengthen this option

## Recommended Actions
Prioritized, concrete, time-bound action items. Not vague "consider X" — specific "do X by Y because Z."
Include a separate "Tech Bets" subsection: which emerging projects to start experimenting with now, even if they're not production-ready.

## Contrarian View
The strongest argument against the consensus/recommended path. What could make all of this wrong?

## Confidence & Gaps
- What we're confident about and why
- What we couldn't determine and what additional research would help
- Key assumptions that should be monitored

## Sources
Consolidated, deduplicated list of all sources across threads.
Phase 4: Save & Deliver
  1. Save the full analysis to 05-knowledge/research/YYYY-MM-DD-[slug].md
  2. If the analysis is long (>3000 words), also create a brief 1-page summary at 05-knowledge/research/YYYY-MM-DD-[slug]-summary.md
  3. Present the Executive Summary + Recommended Actions to the user directly in chat

Show full SKILL.md (409 more words)Show less

Quality Standards

  • No hallucinated sources. Every claim must trace to a real WebSearch/WebFetch result.
  • Recency matters. Prioritize sources from the last 6 months. Flag anything older.
  • Bias awareness. Note when sources have obvious commercial incentives.
  • Specificity over generality. "The testing tools market is $XX.XB and growing at YY% CAGR" beats "the market is growing."
  • Actionability. The output should help the user make a decision, not just understand a topic.
  • Intellectual honesty. If the research is inconclusive, say so. Don't manufacture false confidence.

Example Decomposition

Question: "If generic LLM models get better over time, what's the future for LLM wrapper companies like Katalon or Scout?"

Threads:

  1. Foundation model trajectory — How fast are GPT/Claude/Gemini improving at code understanding, test generation, bug detection? What's the capability curve?
  2. Historical precedent: platform commoditization — What happened to companies built on top of AWS, iOS, Salesforce, etc. when the platform absorbed their features? Who survived and why?
  3. Testing industry structure — Current market map, value chain, where margin lives, what buyers actually pay for
  4. Wrapper company strategies — How are current AI wrapper companies (Jasper, Copy.ai, Cursor, etc.) adapting? What's working?
  5. Enterprise buying behavior — Do enterprises buy "AI" or do they buy "solutions"? What's the procurement reality?
  6. Emerging tech & architectures — What pre-mainstream concepts could reshape the landscape? (e.g., novel agent frameworks, new testing paradigms, computer-use agents, browser automation architectures). Search GitHub repos, arXiv, Twitter/X builder threads, Discord communities, conference talks.
  7. Defensibility analysis — What moats exist for testing-specific AI companies? Data, workflow, integration, brand, switching costs?
  8. Contrarian: wrappers win — Arguments for why vertical AI companies might actually INCREASE in value as models commoditize

Runtime Expectations

  • Phase 1: ~2 minutes (decomposition + user confirmation)
  • Phase 2: ~5-10 minutes (parallel research, longest agent determines total time)
  • Phase 3: ~3-5 minutes (synthesis)
  • Total: ~10-15 minutes for a comprehensive strategic analysis

Error Handling

  • If a research thread returns low-quality results, note this in the synthesis rather than fabricating depth
  • If WebSearch/WebFetch fails for a thread, retry once with alternative search terms, then document the gap
  • The user may interrupt during Phase 2 to redirect or add threads
  • The skill can be run multiple times on related questions — reference previous research files from 05-knowledge/research/

Fallback Behavior

This skill requires WebSearch and WebFetch tools. If these are unavailable:

  • Fall back to vault-only analysis using existing 05-knowledge/ content
  • Clearly state that no live web research was performed
  • Recommend the user run the skill again when web tools are available

© huytieu, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

Just SKILL.md in skills/auto-research of huytieu/COG-second-brain.

Open the folder on GitHubat commit 36ac9d7

Compare with similar skills

Auto Research next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Auto Research compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Auto Research this skillhuytieu/COG-second-brain1.3k—~2.9kAutomated safety check: PassMIT
Scientific Brainstormingspacering-net/codeg3.9k13 repos~2kAutomated safety check: PassMIT
MFA Pipeline Orchestratoraiming-lab/AutoResearchClaw15k—~923Automated safety check: PassMIT
Research RefinezjYao36/Auto-Research-Refine1286 repos~6.9kAutomated safety check: NotesNone
Read GitHubAgentTeam-TaichuAI/ScienceClaw6712 repos~638Automated safety check: PassNone
Web ResearchJuncai22/spring-ai-agent-learning1242 repos~1.1kAutomated safety check: PassApache-2.0

Similar skills

  • Scientific Brainstorming

    spacering-net/codeg

    Creative research ideation and exploration. An agent skill from spacering-net/codeg.

    3.9k GitHub starsUsed in 13 repos~2k tokens
    Research & ScienceAuto-check passed
  • MFA Pipeline Orchestrator

    aiming-lab/AutoResearchClaw

    Runs a metabolic flux analysis from model loading to phenotype prediction and figures by handing work to four sub-agents in sequence.

    15k GitHub stars~923 tokensUpdated 1 mo ago
    Research & ScienceAuto-check passed
  • Research Refine

    zjYao36/Auto-Research-Refine

    Turns a vague research direction into a focused, problem-anchored method plan through up to five review rounds with a second model.

    128 GitHub starsUsed in 6 repos~6.9k tokens
    Research & ScienceAuto-check: notes
  • Read GitHub

    AgentTeam-TaichuAI/ScienceClaw

    Read and search GitHub repository documentation via gitmcp.io MCP service.

    671 GitHub starsUsed in 2 repos~638 tokens
    Research & ScienceAuto-check passed
  • Web Research

    Juncai22/spring-ai-agent-learning

    A skill your agent uses for requests related to web research; it provides a structured approach to conducting comprehensive web research

    124 GitHub starsUsed in 2 repos~1.1k tokens
    Research & ScienceAuto-check passed
  • Deepdive

    Socialpranker/deepdive

    Meta-research под вопрос или решение: веб-поиск, источники, Q&A отчёт с цитатами по файлам для повторного использования.

    372 GitHub stars~5.2k tokensUpdated 6 days ago
    Research & ScienceAuto-check passed

More from huytieu/COG-second-brain

All 37 skills in this repo
  • Voice Baseline

    huytieu/COG-second-brain

    Measure your own writing corpus for the words and sentence shapes you over-use, so an agent writing in your voice stops amplifying your tics into a style.

    1.3k GitHub stars~957 tokensUpdated 8 days ago
    Auto-check passed
  • Daily Journal

    huytieu/COG-second-brain

    A passive daily work journal that Claude keeps FOR you so you never have to write it yourself.

    1.3k GitHub stars~1.3k tokensUpdated 8 days ago
    Auto-check passed
  • Editorial Illustrations

    huytieu/COG-second-brain

    Generate meaning-carrying editorial data-illustrations in the monotykamary / Linear aesthetic (near-black grayscale, Inter display + mono labels, hairline framed figures) with a single coral accent.

    1.3k GitHub stars~2.4k tokensUpdated 8 days ago
    Auto-check passed
  • Loop Engineering

    huytieu/COG-second-brain

    Shared loop-engineering reference for COG skills - the agent loop, deterministic verifiers, termination conditions, in-loop context management, and named patterns.

    1.3k GitHub stars~1.5k tokensUpdated 8 days ago
    Auto-check passed
  • Release Video

    huytieu/COG-second-brain

    Turn a product release (the list of shipped items plus real screen recordings) into a motion recap video and one explained demo per feature, with sound effects tied to on-screen motion and a…

    1.3k GitHub stars~1.7k tokensUpdated 8 days ago
    Auto-check passed
  • Onboarding

    huytieu/COG-second-brain

    Personalize COG for your workflow - creates profile, interests, and watchlist files with guided setup (run this first!)

    1.3k GitHub starsUsed in 1 repo~3.1k tokens
    Auto-check passed

Questions about Auto Research

What does Auto Research do?

Deep strategic research engine — decomposes questions into parallel research threads, spawns multiple agents, and synthesizes into actionable strategic analysis. Auto Research is an agent skill from huytieu/COG-second-brain.

When should I use Auto Research?

Auto Research fits situations like: research & Science work in your project.

How do I install Auto Research in Claude Code?

Run `npx skills add huytieu/COG-second-brain --skill auto-research -a claude-code`. Or copy the skill folder (skills/auto-research in huytieu/COG-second-brain) into .claude/skills/auto-research in your project. Claude Code loads it when a task matches its description.

How do I install Auto Research in Codex?

Run `npx skills add huytieu/COG-second-brain --skill auto-research -a codex`. Or copy the skill folder (skills/auto-research in huytieu/COG-second-brain) into .agents/skills/auto-research in your project. Codex loads it when a task matches its description.

Can I use Auto Research in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add huytieu/COG-second-brain --skill auto-research -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/auto-research, .gemini/skills/auto-research, .github/skills/auto-research and .opencode/skills/auto-research in your project.

What does Auto Research need to run?

SKILL.md names no scripts, command-line tools or credentials: Auto Research is instructions for the agent only.

Does Auto Research access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Auto Research safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Auto Research use?

Auto Research is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Auto Research use?

About 2.9k tokens (SKILL.md is roughly 12k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Auto Research?

Skills that share tags, products or a category with Auto Research: Scientific Brainstorming (spacering-net/codeg, 3.9k stars), MFA Pipeline Orchestrator (aiming-lab/AutoResearchClaw, 15k stars), Research Refine (zjYao36/Auto-Research-Refine, 128 stars) and Read GitHub (AgentTeam-TaichuAI/ScienceClaw, 671 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Auto Research?

huytieu (a GitHub user) maintains it in huytieu/COG-second-brain, which has 1,268 GitHub stars. The repository holds 37 skills in this directory. The repository was last updated on October 2, 2026.

Source: huytieu/COG-second-brain on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.