deep-research

deep-research is a command for Claude Code from ParkerM2/create-claude-workflow. It costs 36 tokens per session (4,982 once invoked), scanned A, original, MIT.

A guided research command that investigates a topic in stages and produces a report with sources and analysis. It includes checkpoints where the user can steer the investigation.

In plain words
What is it for?
Use it to study technologies, compare approaches, investigate how other tools work, or gather context before creating an implementation plan or coordinating feature work.
Why use it?
It provides checked and cross-referenced information before an important technical decision. It helps when trying an approach immediately would be risky or when the codebase needs deeper understanding.

Command for Claude Code

Part of the claude-workflow plugin — 10 skills, 25 commands, 3 agents, 6 hooks shipped together

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.

agentmods
npx agentmods add commands/parkerm2/create-claude-workflow/deep-research
Clone the repo
git clone --depth 1 https://github.com/ParkerM2/create-claude-workflow

Made for: Claude Code.

Or install claude-workflow, the plugin that ships this one along with the rest of its 10 skills, 25 commands, 3 agents, 6 hooks.

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/commands/parkerm2/create-claude-workflow/deep-research.svg)](https://agentmods.dev/commands/parkerm2/create-claude-workflow/deep-research)
Your own site
<a href="https://agentmods.dev/commands/parkerm2/create-claude-workflow/deep-research"><img src="https://agentmods.dev/badge/commands/parkerm2/create-claude-workflow/deep-research.svg" alt="Measured on agentmods" height="20"></a>
Per session 36 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 4,982 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
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 $0.00036 $0.04982
Opus 5 $0.00018 $0.02491
Sonnet 5 $0.00007 $0.00996
Haiku 4.5 $0.00004 $0.00498

Measured 3d ago against content hash b7c51f6269ca, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, 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 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.

.claude/commands/deep-research.md · 578 lines

How it starts

The opening of the file, as written. The whole thing — 578 lines — stays where its author put it; the contents beside it link to each section on GitHub.

/deep-research — Conversational Deep Research

Deeply research a topic before making decisions. Works through phases with user checkpoints between each — you control the direction. Each layer of research is cross-referenced against the goal, prior findings, and codebase reality before presenting. Produces a research report, not a design doc. Use /new-plan after to turn findings into an implementation plan.


When to Use

  • Exploring a new technology, pattern, or approach before committing to it
  • Investigating how other tools/frameworks solve a problem you're facing
  • Understanding the current codebase deeply before planning changes
  • Comparing multiple approaches with trade-offs before choosing one
  • Gathering context that will inform a future /new-plan or /agent-team
  • Any question where "just try it" is too risky and you need data first

When NOT to Use

  • You already know what to build — use /new-plan or /agent-team directly
  • Simple factual question — just ask directly without invoking a skill

Core Principle: Research Before Presenting

Never present raw findings directly. Every piece of information goes through a validation layer before the user sees it:

Raw finding → Cross-reference against goal → Validate against prior research → Check against codebase → Assess confidence → THEN present

If you can't validate a finding, say so explicitly with the reason. Don't present unvalidated information as fact.


How It Works

Phase 1: Scope           → Define question, dimensions, success criteria
     ↓ checkpoint
Phase 2: Explore          → Broad research across all dimensions
     ↓ internal validation (DO NOT present raw findings)
Phase 2.5: First Pass     → Cross-reference, filter, assess confidence
     ↓ checkpoint (present validated findings only)
Phase 3: Analyze          → Deep dive on user-selected threads (ITERATIVE)
     ↓ validate each round against goal + prior context
     ↓ checkpoint per round (user can loop or advance)
Phase 4: Synthesize       → Final validation pass, compile report
     ↓ checkpoint (user can refine)

Read the full file on GitHub · 578 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. 3d ago First seen · 578 lines · 36 tokens per session scan A b7c51f6269ca

Subscribe to this mod's changes

deep-research is a command published in the GitHub repository ParkerM2/create-claude-workflow (4 stars, last pushed 5mo ago), licensed MIT. It adds 36 tokens to every session and 4,982 once invoked, about $0.0002 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 2026-08-31.