Deep Research

What it does

When your agent uses it

• Whats the latest on
• Asks about something they want to understand quickly but well

Example prompts

• “research,”
• “deep dive on,”
• “tell me about,”
• “/deep-research”

Workflow steps

5 steps, taken from the first numbered list in SKILL.md.

1. 1Topic — what specifically?
2. 2Why now — what's prompting the research? (Meeting, decision, content piece, just curious?)
3. 3Depth — quick (10 min of effort) or deep (30+)?
4. 4Format preference — straight brief, comparison table, decision-oriented?
5. 5What you already know — so the brief doesn't waste space repeating it

What it can do on your machine

Read from SKILL.md and the folder at commit fdbf39b. 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

Deep Research loads about 1.5k tokens when it runs. Until then it costs about 136 tokens; SKILL.md has 620 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it

~136

When it runs · the whole SKILL.md, loaded when a task matches

~1.5k

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

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 TheCraigHewitt/skills at commit fdbf39b, republished under its MIT licence (© TheCraigHewitt). 620 words, ~1,459 tokens.

name

deep-research

description

Research a topic deeply across multiple sources and produce a sourced, bite-sized brief the user can read in 5 minutes. Use this whenever the user says 'research,' 'deep dive on,' 'tell me about,' 'what's the latest on,' 'do a brief on,' 'TL;DR of [topic],' 'catch me up on,' or asks about something they want to understand quickly but well. Works for AI tools, companies, people, technical concepts, market trends, news events, scientific topics. Output is a markdown brief with sources, pros/cons, and your read on what matters.

# [Topic]

**Date:** YYYY-MM-DD
**Depth:** Quick / Standard / Deep
**Why I researched this:** [one line]

## TL;DR
3–4 sentences. The headline answer.

## What it is / who they are / what happened
The core facts. 1–3 paragraphs. Plain English.

## The state of play
What's happening with this right now. Recent developments, current debates, where the action is.

## Pros / strengths / what's working
Bullet list.

## Cons / weaknesses / open questions
Bullet list. Be honest — don't hedge.

## Compared to alternatives
For tools/companies/products: brief comparison to 2-3 alternatives. For concepts: how this relates to adjacent ideas.

## What the experts disagree on
The honest disagreement in the space. Cite which expert says what.

## My read
Your synthesis — what you think the user should take away. One paragraph. Be willing to render a verdict, especially if the user is researching this to make a decision.

## Sources
Numbered list with title, link, and a one-line note on why this source matters. Aim for 5-10 strong sources rather than 30 weak ones.

## Open questions / gaps
What I couldn't find out or what would require deeper investigation.

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

Files

SKILL.md and 1 other file in cowork/deep-research of TheCraigHewitt/skills.

• SKILL.md
• evals/evals.json

Open the folder on GitHubat commit fdbf39b