A skill for Claude Code and Codex · Research and strategy
deep-search-expert makes Claude Code or Codex research like an analyst: it plans its angles before searching, cross-checks official docs, forums, GitHub, and video, and only concludes with evidence tied to every claim. It's the one I use when the decision costs real money and a three-link summary won't cut it.
Solo skill: unlocks right away, no guarantee after download. Purchase terms.
Or get all 25 skills plus 9 bonuses in the Supreme Pack for R$ 497.
The problem
The problem isn't that AI can't find information. It's that it stops at the first result that looks good and hands you a confident summary without saying where each sentence came from. Three symptoms you've already seen, and what changes once the skill is running the show.
| Without the skill | With deep-search-expert |
|---|---|
| You ask a question and it answers with total confidence, except half the claims have no source and the other half came from an old article that's no longer accurate. | Every key conclusion ships with the sources that support it and the ones that contest it, dated, and anything unproven shows up as a gap, not as fact. |
| The research stops at the first Google result and marketing copy, while the real answer was sitting in a GitHub issue, a forum thread, or a 40-minute video. | It splits the topic into independent angles, official sources, secondary coverage, community, GitHub, and video, covers whichever ones fit the question, and logs the ones it skipped. |
| Every time the topic comes back up, you research it from scratch, burning credits, time, and context, and land on a conclusion slightly different from last time. | Before searching, it checks what's already been researched and reuses it, leaving behind a manifest with the plan, sources, and claims you can reopen next round. |
What you get out of it
In the end you get research you can audit, not pretty copy you just have to take on faith.
every relevant claim with a confidence level and the evidence for and against it
primary sources, secondary reporting, opinion, and inference each kept in their own lane
goal, angles, sources with links and dates, and whatever gaps stayed open
when the topic touches code, the GitHub project gets analyzed without running any of it
How it works
An explicit request starts right away. When the request is vague, it states which angles it plans to cover and checks in before burning time and tokens.
Sets the goal, recency window, angles, source budget, and the minimum evidence per claim, then reuses whatever's already been researched on the same topic.
Searches official docs, articles, forums, GitHub, and video, pulling captions or a transcript when needed. If a search tool fails or credits run out, it switches paths instead of stalling.
Every important claim becomes a record with the sources that back it and the ones that contest it. A social media post only counts as fact once another source confirms it.
Wraps up when the sources converge or the budget runs out, and delivers the synthesis with a confidence score per claim and the gaps spelled out.
Before you buy
Video without captions can be transcribed in the cloud or on your own machine. Both paths are optional, and the skill checks with you first when the video is long.
Its limits
The skill delivers organized evidence and a confidence level for every conclusion, but the decision is still yours: when sources disagree, it shows you the disagreement instead of picking a side.
Deep research burns real time and real tokens, so for a quick one-line question it's overkill and the agent's regular search does the job. It pays off on benchmarks, competitor analysis, and expensive technical decisions.
Content behind a login only gets in with your own authorized session, and that kind of link never gets sent to an outside service.
What you receive
The price
Just deep-search-expert
R$97one time payment
Bought it solo and now want it all? Message me on WhatsApp and I will deduct what you already paid.
Supreme Pack
R$ 1.665R$497or 10x on a card
Buying all 25 solo adds up to R$ 1.665. The pack pays for itself by the ninth skill. See everything the Supreme Pack includes
01 · Research and strategy
Before you ask
Regular search answers with whatever it finds first. The skill plans its angles, cross-checks different types of sources, logs each one with a date, and ties every conclusion to the evidence for and against it. In the end you know what's fact, what's opinion, and what's still unproven.
No. With a key, page collection is more thorough; without one, the skill falls back on Claude Code's or Codex's native search and reading, following the exact same method of plan, evidence, and gaps.
That's one of its main uses, along with tool benchmarking and technical decisions. It cross-checks the official site, reviews, forums, and GitHub, and separates what the company claims from what users actually report.
It depends on the source budget set in the plan. With just a few angles it wraps up in minutes; a broad investigation takes longer and burns more tokens, which is why it checks in before starting whenever the request isn't explicit.
Yes, the license covers your work for third parties. What's off limits is reselling or redistributing the skill.
One-time payment, 12 months of updates. No refunds after download; the Supreme Pack comes with a 7-day guarantee.
