Deep research
Plug-and-play skills and prompts for every AI coding agent
npx -y skills add Amey-Thakur/AI-SKILLS --skill deep-researchAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
2 things to look at
- 19 days oldThe repository was created 19 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
- 4 stars4 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Run a thorough multi-source research investigation end to end: plan, search broadly, verify, and synthesize a cited answer. Use when a question needs deep, autonomous research across many sources, not a quick lookup.
SKILL.md
3.3 KB, as published. Nobody here has run it
Deep research
Deep research is the full loop from a hard question to a trustworthy, cited answer: plan the investigation, gather from many sources, verify what you find, and synthesize it into something more than the sum of the pages. It is the method an agent (or a person) follows to research autonomously without cutting corners.
Method
- Plan before searching. Sharpen the question, decompose it into answerable sub-questions, and decide what evidence would settle each and where it lives (see research-planning). A deep investigation without a plan becomes a pile of tangents.
- Search broadly, from multiple angles. Cover each sub-question from several search angles and source types, so you are not captured by the first framing or the most popular result (see web-research). Cast wide first (find the landscape, the key sources, the disagreements), then go deep on what matters. Deliberately seek disconfirming evidence and the strongest opposing view, not just support.
- Go deep on primary sources. Read the actual papers, docs, data, and original statements for the load-bearing claims, not summaries of them (see reading-papers, source-evaluation). The depth that distinguishes deep research is following claims to their origin and understanding them, not skimming aggregators.
- Verify as you go. Evaluate every source and corroborate every important claim across independent origins before it enters your conclusion (see source-evaluation, fact-checking). Do not let an unverified claim propagate into the synthesis; mark what is confirmed, what is contested, and what is unknown.
- Synthesize, do not just collect. The value is in connecting the findings: reconciling or surfacing the disagreements between sources, noting the consensus and the outliers, and answering the actual question rather than dumping what you found (see research-synthesis). A list of quotes is not research; the reasoned answer that accounts for them is.
- Deliver with citations and honest confidence. Present the answer with its sources so it can be checked, distinguish well-supported conclusions from tentative ones, and state the open questions and what would resolve them (see writing-with-evidence). Know when to stop: when new sources stop changing the answer (saturation) or the decision is adequately supported.
Boundaries
- Deep research is expensive; reserve it for questions whose stakes justify it. Most questions need a quick lookup and a source check, not the full loop (see research-planning's depth budget).
- Thoroughness is not the appearance of it: many low-quality sources are not depth, and a long report is not a good one. Depth is verified primary evidence and real synthesis, not volume (see the completeness-vs-padding tension in research-synthesis).
- For an autonomous agent, the loop must include verification and grounding at every step; an agent that searches and synthesizes without fact-checking confidently launders errors into a citation-shaped answer (see fact-checking, llm-guardrails).