learn
by tw93Learn is a research skill that turns unfamiliar topics, source bundles, and collected notes into a coherent, publish-ready output. It helps with deep dives, source compilation, explanation, and structured synthesis for web research and other multi-source tasks. Best when you need one strong reference from many inputs, not a quick lookup.
This skill scores 78/100, which means it is a solid listing candidate for directory users who need a structured multi-source research workflow. The repository gives enough trigger guidance and operational detail to help an agent choose and use it more reliably than a generic prompt, though install-time users should still expect some missing ecosystem support and a few workflow gaps.
- Clear intended use: multi-source deep research, source compilation, and publish-ready synthesis are explicitly named in the description and when_to_use fields.
- Strong operational framing: the skill lays out a six-phase research workflow and distinguishes `/learn` from `/read`, which helps agents trigger it correctly.
- Good progressive disclosure: the body is substantial, with many headings and constraint sections, suggesting a real workflow rather than a placeholder.
- No install command or companion files are present, so users may need to infer setup and adjacent skill dependencies from the text alone.
- The excerpt shows pre-check and mode-selection logic, but support files are absent and the repository exposes no scripts/resources to validate or automate the workflow.
Overview of learn skill
What learn does
The learn skill is a research workflow for turning scattered material into a coherent, publish-ready output. It is the right learn skill when you need to study a topic, compile sources, and produce one structured answer instead of a quick lookup.
Best-fit use cases
Use learn for Web Research when the job is to synthesize multiple pages, notes, or documents into a single reference, article, brief, or explainer. It fits “deep dive,” “整理一下,” “研究一下,” and “compile sources” workflows better than one-off summarization.
Why it stands out
The main value of learn is structure: it helps organize raw inputs, explain them, and turn them into something usable. It is strongest when the user already has a topic or source bundle but needs better synthesis, clearer framing, and a cleaner final artifact.
How to Use learn skill
Install and enable learn
Install with npx skills add tw93/Waza --skill learn. The repository does not include helper scripts, so the key installation check is simply that skills/learn/SKILL.md is available and the skill is loaded in your environment.
Start from the right prompt
A good learn usage prompt names the topic, target format, source set, and audience. For example: “Use learn to turn these five sources into a concise briefing for a product manager, highlighting risks, definitions, and open questions.” This gives the skill enough context to research and structure the output well.
Read these files first
Start with skills/learn/SKILL.md, then review any repo-level guidance such as README.md or AGENTS.md if they exist in your environment. In this repo, SKILL.md is the main source of truth, so understanding its phases and boundaries is the fastest way to use the learn guide correctly.
Work with the workflow, not against it
learn is designed for multi-source research. It is not ideal for a single URL that only needs a fetch or a simple answer. If your task is borderline, describe whether you need collection, comparison, explanation, or final publication quality so the skill can apply the right depth.
learn skill FAQ
Is learn only for web research?
No. learn is useful for learn for Web Research, but it is broader than that. It also works for mixed source bundles, notes, documents, and research notes that need to be unified into one result.
When should I not use learn?
Do not use learn for a single quick lookup or a one-page summary that does not require synthesis. If you only need to fetch and read one page, a simpler skill or direct prompt is usually faster.
Do I need to be an expert to use it?
No. The skill is beginner-friendly if you can describe the topic and the desired output. The better your source list and audience definition, the better the result, but you do not need to know the domain deeply upfront.
How is it different from a normal prompt?
A normal prompt often asks for an answer. learn is closer to a structured research pass: it expects source gathering, organization, explanation, and a final output that can stand on its own. That makes it more reliable for messy topics and multi-source tasks.
How to Improve learn skill
Give it a sharper target
The most effective way to improve learn output is to specify the end product: memo, article, explainer, FAQ, comparison table, or decision brief. Include the audience, desired tone, and what must be true in the final answer.
Provide better source boundaries
learn works best when you tell it which sources matter and what to ignore. For example, name the docs, URLs, or notes to prioritize, and say whether you want only primary sources, only recent material, or both. This reduces drift and shallow synthesis.
Watch for common failure modes
The main failure modes are vague topic scope, source overload, and outputs that summarize without concluding. If the first pass feels broad, narrow the question, add constraints, or ask for a second pass that focuses on tradeoffs, definitions, or action items.
Iterate after the first output
Use the first result to spot missing angles: definitions, disagreements, edge cases, or practical implications. Then ask learn to revise with those gaps filled. That is usually more effective than asking for a longer first draft.