[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"article-6-part-prompt-scoring-turns-vague-prompts-into-usable-ones-en":3,"article-related-6-part-prompt-scoring-turns-vague-prompts-into-usable-ones-en":30,"series-tools-2d7d46e9-e387-449c-97d9-2e99eddb3496":83},{"id":4,"slug":5,"title":6,"content":7,"summary":8,"source":9,"source_url":10,"author":11,"image_url":12,"cover_image":12,"category":13,"language":14,"translated_content":11,"related_article_id":15,"keywords":16,"key_takeaways":22,"views":26,"created_at":27,"published_at":28,"topic_cluster_id":29},"2d7d46e9-e387-449c-97d9-2e99eddb3496","6-part-prompt-scoring-turns-vague-prompts-into-usable-ones-en","6-part prompt scoring turns vague prompts into usable ones","\u003Cp data-speakable=\"summary\">A six-part prompt checklist you can copy before sending anything to an AI model.\u003C\u002Fp>\u003Cp>I've been writing prompts for \u003Ca href=\"\u002Ftag\u002Fchatgpt\">ChatGPT\u003C\u002Fa>, \u003Ca href=\"\u002Ftag\u002Fclaude\">Claude\u003C\u002Fa>, and \u003Ca href=\"\u002Ftag\u002Fgemini\">Gemini\u003C\u002Fa> for a while now, and I kept hitting the same annoying wall: the model would sound confident, but the output would still be off. Not wildly off. Worse. Close enough to waste my time. I'd ask for a rewrite and get something polished but useless. I'd ask for a plan and get a list that looked fine until I tried to use it. The problem wasn't the model. It was me being sloppy with the prompt.\u003C\u002Fp>\u003Cp>What finally snapped this into focus was a guide from \u003Ca href=\"https:\u002F\u002Fwww.moreonlinetools.com\u002Fen\u002Fblog\u002Fprompt-engineering-complete-guide\u002F\">MoreOnlineTools\u003C\u002Fa>. They took the mushy idea of \"good prompting\" and broke it into six criteria: clarity, role, context, format, constraints, and examples. That's the part I care about. Not the theory, not the hype. A practical checklist that tells me why a prompt failed and how to fix it before I waste another round trip with the model.\u003C\u002Fp>\u003Cp>And yes, I know everyone has a prompt framework now. Most of them are just vibes with bullet points. This one is better because it gives me something I can score, inspect, and apply immediately. I can look at a prompt and say, \"You're vague here, you're missing context there, and you're begging the model to guess the format.\" That kind of diagnosis is useful.\u003C\u002Fp>\u003Ch2>Stop asking the model to read your mind\u003C\u002Fh2>\u003Cblockquote>\"AI models are not mind readers. They respond to exactly what you write. The more precisely you describe what you want — the role, the context, the format, the constraints — the better the result.\"\u003C\u002Fblockquote>\u003Cp>That line is the whole game. What this actually means is that prompt quality is usually a specification problem, not an intelligence problem. If the output feels generic, I don't blame the model first. I check whether I gave it enough shape to work with.\u003C\u002Fp>\n\u003Cfigure class=\"my-6\">\u003Cimg src=\"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1781151511274-pig7.png\" alt=\"6-part prompt scoring turns vague prompts into usable ones\" class=\"rounded-xl w-full\" loading=\"lazy\" \u002F>\u003C\u002Ffigure>\n\u003Cp>I ran into this constantly when I used AI for drafts. I'd say, \"Write me a landing page.\" The model would produce something technically correct and completely forgettable. Then I'd complain that it was bland. But bland is what you get when you ask for a thing without telling it who it's for, what the offer is, what tone to use, and what result you want.\u003C\u002Fp>\u003Cp>How to apply it: before you hit enter, ask yourself if the prompt contains all four of these things: what the task is, who it's for, what format you want back, and what should be avoided. If one of those is missing, the model has room to wander.\u003C\u002Fp>\u003Cul>\u003Cli>State the task in plain language.\u003C\u002Fli>\u003Cli>State the audience or use case.\u003C\u002Fli>\u003Cli>State the output shape.\u003C\u002Fli>\u003Cli>State the boundaries.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>This is boring, which is exactly why it works. AI doesn't reward cleverness as much as it rewards specificity. I wish that weren't true, because I enjoy a good elegant prompt. But the ugly, explicit one usually wins.\u003C\u002Fp>\u003Ch2>Clarity is the cheapest upgrade you can make\u003C\u002Fh2>\u003Cblockquote>\"Replace vague words with precise descriptions.\"\u003C\u002Fblockquote>\u003Cp>Clarity gets the biggest weight in the source guide, and that makes sense. If the request is fuzzy, everything downstream gets fuzzy too. A prompt like \"give me some ideas\" is basically a coin flip. A prompt like \"give me five headline ideas for a blog post about reducing churn in subscription businesses, each under 60 characters and with urgency\" gives the model something to actually solve.\u003C\u002Fp>\u003Cp>What this actually means is that the model needs a target, not a mood. \"Help me with marketing\" is a mood. \"Write a 300-word LinkedIn post about email open rates for B2B SaaS\" is a target. One invites improvisation. The other creates useful pressure.\u003C\u002Fp>\u003Cp>I used to think being specific would make prompts too rigid. Sometimes that's true. But in practice, vague prompts don't leave room for creativity; they leave room for confusion. The model is not going to infer your hidden preferences with any reliability. It will fill the gaps with whatever is statistically common.\u003C\u002Fp>\u003Cp>How to apply it: whenever you write a prompt, force yourself to add numbers, nouns, and audience details. Numbers help. \"Short\" is slippery. \"Under 120 words\" is usable. \"For founders\" is vague. \"For first-time SaaS founders raising seed funding\" is better.\u003C\u002Fp>\u003Cul>\u003Cli>Use measurable limits when possible.\u003C\u002Fli>\u003Cli>Replace abstract words with concrete output targets.\u003C\u002Fli>\u003Cli>Include domain details the model would otherwise guess.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>This is the part most people skip because it feels tedious. It is tedious. It also saves you three follow-up prompts and a bad edit pass.\u003C\u002Fp>\u003Ch2>Role is how you stop getting generic answers\u003C\u002Fh2>\u003Cblockquote>\"Assigning the AI a specific role dramatically shifts the tone, depth, and perspective of its response.\"\u003C\u002Fblockquote>\u003Cp>The source guide recommends role prompting for good reason. Telling the model who it is changes the frame it uses to answer. \"Explain quantum computing\" and \"You are a physics professor explaining quantum computing to a first-year undergraduate\" are not the same request, even though they point at the same topic.\u003C\u002Fp>\n\u003Cfigure class=\"my-6\">\u003Cimg src=\"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1781151504987-lz15.png\" alt=\"6-part prompt scoring turns vague prompts into usable ones\" class=\"rounded-xl w-full\" loading=\"lazy\" \u002F>\u003C\u002Ffigure>\n\u003Cp>What this actually means is that role is a shortcut for style and judgment. If I want a brutally honest editor, I say that. If I want a senior DevOps engineer, I say that. If I want a Socratic tutor, I say that too. The model doesn't become that person, obviously, but it does start producing from that angle.\u003C\u002Fp>\u003Cp>I use this most when the first answer is too polished and not opinionated enough. That happens a lot with marketing copy. If I say, \"review this homepage copy,\" I get soft feedback. If I say, \"you are a senior conversion copywriter and your job is to cut anything weak,\" the answer gets sharper immediately.\u003C\u002Fp>\u003Cp>How to apply it: choose a role that has a real point of view, not just a title. \"Writer\" is weak. \"Technical editor who hates filler\" is better. The role should imply how the model should think, not just what costume it should wear.\u003C\u002Fp>\u003Cp>Useful roles I keep reusing:\u003C\u002Fp>\u003Cul>\u003Cli>Senior editor\u003C\u002Fli>\u003Cli>Staff engineer\u003C\u002Fli>\u003Cli>Data analyst\u003C\u002Fli>\u003Cli>Product manager\u003C\u002Fli>\u003Cli>Brutally honest reviewer\u003C\u002Fli>\u003C\u002Ful>\u003Cp>That list is not magic. It just works because each role carries expectations. The model is much better when those expectations are explicit.\u003C\u002Fp>\u003Ch2>Context is the part that saves you from rework\u003C\u002Fh2>\u003Cblockquote>\"The AI does not know your situation unless you tell it.\"\u003C\u002Fblockquote>\u003Cp>This one hurts because it is where most of my bad prompts fail. I used to assume the model would infer the backstory from the question. It won't. If the audience matters, say so. If the business goal matters, say so. If you've already tried something and it didn't work, say that too.\u003C\u002Fp>\u003Cp>What this actually means is that context is the difference between a usable answer and a technically correct answer. A client email is not just an email. It's a follow-up after 10 days of silence on a $15,000 proposal. That detail changes the tone, the urgency, and the wording.\u003C\u002Fp>\u003Cp>I hit this when I asked for a project summary and got something that read like a generic internal memo. The model hadn't been told whether the reader was technical, executive, customer-facing, or brand new to the topic. So it guessed. Badly. Once I added audience, goal, and prior attempts, the output got much more useful.\u003C\u002Fp>\u003Cp>How to apply it: include the minimum background required for the model to make a good decision. Not the whole history. Just the part that changes the answer.\u003C\u002Fp>\u003Cul>\u003Cli>Who is this for?\u003C\u002Fli>\u003Cli>What is the goal?\u003C\u002Fli>\u003Cli>What has already been tried?\u003C\u002Fli>\u003Cli>What constraints are already in place?\u003C\u002Fli>\u003C\u002Ful>\u003Cp>If the model needs to choose between multiple valid directions, context is what tells it which one matters.\u003C\u002Fp>\u003Ch2>Format is not decoration, it's part of the task\u003C\u002Fh2>\u003Cblockquote>\"Requesting a specific format saves editing time.\"\u003C\u002Fblockquote>\u003Cp>I think people underrate format because they treat it like presentation. It's not. It's instruction. If I want a Markdown table, numbered list, JSON object, or H2\u002FH3 structure, I need to say that up front. Otherwise the model will hand me a wall of text and act like it did its job.\u003C\u002Fp>\u003Cp>What this actually means is that format is a constraint on thinking. A table forces comparison. A list forces ordering. JSON forces structure. Headings force hierarchy. The output shape changes the quality of the reasoning because it changes what the model has to fill in.\u003C\u002Fp>\u003Cp>I use format prompts when I know the output has to be pasted somewhere else. For example, when I need a structured summary for a doc, I ask for bullets with fixed labels. When I need \u003Ca href=\"\u002Ftag\u002Fcode-review\">code review\u003C\u002Fa> notes, I ask for a table with severity, location, issue, and fix. That saves me from reformatting a blob of text by hand.\u003C\u002Fp>\u003Cp>How to apply it: specify both the container and the fields. Don't just say \"format as a table.\" Say what the columns should be. Don't just say \"write JSON.\" Say which keys you need.\u003C\u002Fp>\u003Cul>\u003Cli>Use numbered lists for ordered steps.\u003C\u002Fli>\u003Cli>Use tables for comparisons and audits.\u003C\u002Fli>\u003Cli>Use JSON when another tool will consume the output.\u003C\u002Fli>\u003Cli>Use headings when the result will be read by humans in a doc.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>This is one of the easiest prompt upgrades because it doesn't require better thinking, just more explicit thinking.\u003C\u002Fp>\u003Ch2>Constraints keep the model from wandering off\u003C\u002Fh2>\u003Cblockquote>\"Constraints dramatically reduce the chance of the AI going off in an unhelpful direction.\"\u003C\u002Fblockquote>\u003Cp>Constraints are the guardrails. They tell the model what not to do, how long to stay, and which options are off limits. I used to skip these because I thought they made prompts too fussy. Then I kept getting answers that were technically fine and operationally annoying.\u003C\u002Fp>\u003Cp>What this actually means is that a good prompt is not just a request. It's a request with boundaries. If I need a response under 200 words, say it. If I don't want jargon, say it. If I only want free tools, say it. Otherwise the model will happily ignore your unstated preferences.\u003C\u002Fp>\u003Cp>I use constraints most when the output has a real use case. If the answer is going into a Slack message, I want it short. If it's going into a client email, I want the tone warm but not mushy. If it's for a non-technical audience, I want plain language and no acronyms unless I define them. The more the output has to fit a real container, the more constraints matter.\u003C\u002Fp>\u003Cp>How to apply it: write at least one positive constraint and one negative constraint. Positive means what must be included. Negative means what must be avoided.\u003C\u002Fp>\u003Cul>\u003Cli>Keep it under 150 words.\u003C\u002Fli>\u003Cli>Do not use jargon.\u003C\u002Fli>\u003Cli>Do not recommend paid tools.\u003C\u002Fli>\u003Cli>Include one example.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>That little bit of pressure makes the model behave. Without it, the answer can sprawl into something you don't want to edit.\u003C\u002Fp>\u003Ch2>Examples are the fastest way to teach style\u003C\u002Fh2>\u003Cblockquote>\"Even one good example can transform the quality of the output.\"\u003C\u002Fblockquote>\u003Cp>Few-shot prompting is the part of the guide that feels like cheating, because it kind of is. If you show the model one or two examples of the pattern you want, it starts matching that pattern instead of guessing from scratch. This is especially useful when the task is classification, rewriting, or any output with a consistent house style.\u003C\u002Fp>\u003Cp>What this actually means is that examples compress ambiguity. A written description of style is always a little fuzzy. A sample output is not. The model can infer structure, tone, and level of detail much more reliably from examples than from adjectives.\u003C\u002Fp>\u003Cp>I use examples when I need repeatable output. If I want product descriptions, I give one good sample. If I want sentiment labels, I give a few labeled examples. If I want a summary in a specific voice, I paste a short model example and say \"write the next one like this.\" That usually gets me closer than a long explanation ever would.\u003C\u002Fp>\u003Cp>How to apply it: use examples when the desired output has a recognizable pattern. You don't need ten. Often two or three are enough. The point is to show the model the shape of the answer, not to exhaustively document it.\u003C\u002Fp>\u003Cp>One caution: examples can also lock in bad habits if your sample is sloppy. So yes, the example matters. If your sample is weak, the model will imitate the weakness too.\u003C\u002Fp>\u003Ch2>The scoring mindset is the real trick\u003C\u002Fh2>\u003Cp>MoreOnlineTools says its \u003Ca href=\"https:\u002F\u002Fwww.moreonlinetools.com\u002Fen\u002Fblog\u002Fprompt-engineering-complete-guide\u002F\">free AI Prompt Optimizer\u003C\u002Fa> scores prompts across the six criteria and gives improvement suggestions. I like that idea more than I expected to. Not because scoring is glamorous. It isn't. Because scoring gives you a repeatable way to spot what's missing before you send the prompt.\u003C\u002Fp>\u003Cp>What this actually means is that prompt writing becomes less mystical and more like editing. I can look at a prompt and ask: is it clear, does it define a role, does it include context, does it specify format, does it set constraints, and does it use examples where needed? That's a better workflow than just hoping my instinct is good.\u003C\u002Fp>\u003Cp>I also think this helps teams. If two people write prompts differently, a shared checklist gives them a common standard. Otherwise one person thinks the prompt is \"good enough\" because it sounds smart, while another person is staring at a vague mess and wondering why the output is unusable.\u003C\u002Fp>\u003Cp>How to apply it: treat your prompts like drafts. Score them before use. If you can't score them, you're probably relying on vibes. And vibes are expensive when the model burns a minute producing something you can't use.\u003C\u002Fp>\u003Ch2>The template you can copy\u003C\u002Fh2>\u003Cpre>\u003Ccode>[Persona] You are a [role\u002Fexpert] with experience in [domain].\n\n[Task] Your task is to [specific action] for [audience\u002Fuse case].\n\n[Context] Here is the background you need:\n- Audience: [who this is for]\n- Goal: [what success looks like]\n- Situation: [what has already happened]\n- Constraints: [anything the model must respect]\n\n[Clarity] Be specific about [topic], and focus on [subtopic].\n\n[Format] Present the result as [Markdown table \u002F numbered list \u002F JSON \u002F headings \u002F email \u002F checklist].\nInclude these fields or sections:\n- [field 1]\n- [field 2]\n- [field 3]\n\n[Constraints]\n- Keep the response under [word count \u002F length]\n- Do not use [jargon \u002F filler \u002F paid tools \u002F irrelevant details]\n- Do not include [anything you want excluded]\n\n[Examples]\nExample input: [paste a sample input]\nExample output: [paste a strong sample output]\n\nNow produce the answer for:\n[paste your actual request]\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>Here's the version I actually use in practice when I want to score a prompt before sending it:\u003C\u002Fp>\u003Cpre>\u003Ccode>Prompt review checklist:\n1. Is the request clear and specific?\n2. Is the role defined?\n3. Is enough context included?\n4. Is the output format explicit?\n5. Are constraints stated?\n6. Are examples needed, and if so, are they good?\n\nRewrite the prompt to improve the weakest items first.\nReturn:\n- Score from 0 to 100\n- Weak areas\n- Specific fixes\n- A revised prompt I can copy\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>That second block is the one I'd hand to an AI assistant when I want it to audit my own prompt. It's simple, direct, and hard to misunderstand. Which, annoyingly, is exactly why it works.\u003C\u002Fp>\u003Cp>Source attribution: this breakdown is based on the MoreOnlineTools guide at \u003Ca href=\"https:\u002F\u002Fwww.moreonlinetools.com\u002Fen\u002Fblog\u002Fprompt-engineering-complete-guide\u002F\">https:\u002F\u002Fwww.moreonlinetools.com\u002Fen\u002Fblog\u002Fprompt-engineering-complete-guide\u002F\u003C\u002Fa>. I added the editorial framing, examples, and copy-ready template; the six-part structure and tool description come from the original article.\u003C\u002Fp>","I break down a six-part prompt checklist and turn it into a copy-paste template you can use before sending prompts.","www.moreonlinetools.com","https:\u002F\u002Fwww.moreonlinetools.com\u002Fen\u002Fblog\u002Fprompt-engineering-complete-guide\u002F",null,"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1781151511274-pig7.png","tools","en","dc0b38ea-3f3c-4ec3-83e8-b26692aa54c4",[17,18,19,20,21],"prompt engineering","prompt optimization","LLM prompts","few-shot prompting","AI workflow",[23,24,25],"The best prompts are specific about task, audience, format, and limits.","Role, context, and examples change output quality more than clever wording.","A scoring checklist makes prompt writing repeatable instead of guessy.",0,"2026-06-11T04:18:00.860733+00:00","2026-06-11T04:18:00.854+00:00","a7343b93-37cc-4634-a2bc-707f6275bdb6",{"tags":31,"relatedLang":42,"relatedPosts":46},[32,34,36,38,40],{"name":19,"slug":33},"llm-prompts",{"name":17,"slug":35},"prompt-engineering",{"name":18,"slug":37},"prompt-optimization",{"name":20,"slug":39},"few-shot-prompting",{"name":21,"slug":41},"ai-workflow",{"id":15,"slug":43,"title":44,"language":45},"six-part-prompt-scoring-turns-vague-prompts-into-usable-ones-zh","六項評分把模糊提示變可用","zh",[47,53,59,65,71,77],{"id":48,"slug":49,"title":50,"cover_image":51,"image_url":51,"created_at":52,"category":13},"bc6e4805-34c1-456c-9628-2a1746db6c21","deep-research-prompt-framework-ai-reports-en","Deep Research Prompt Framework for Better AI Reports","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1781155090319-y5q2.png","2026-06-11T05:17:40.460525+00:00",{"id":54,"slug":55,"title":56,"cover_image":57,"image_url":57,"created_at":58,"category":13},"d8f6f103-5502-4fcc-862d-5271e1cc3d69","last30days-skill-best-reason-stop-trusting-search-alone-en","last30days-skill is the best reason to stop trusting search alone","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1781122670367-luxf.png","2026-06-10T20:17:22.707907+00:00",{"id":60,"slug":61,"title":62,"cover_image":63,"image_url":63,"created_at":64,"category":13},"57a4012c-5884-47f1-babd-aa193a10468e","15-ai-coding-assistant-tools-2026-en","15 AI Coding Assistant Tools for 2026","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1781114586004-3e6d.png","2026-06-10T18:02:27.929561+00:00",{"id":66,"slug":67,"title":68,"cover_image":69,"image_url":69,"created_at":70,"category":13},"908bd6d7-ba5a-40f0-8000-e785fd1372f5","cuda-oxide-rust-ptx-kernels-en","cuda-oxide turns Rust into PTX kernels","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1781110153405-wqt4.png","2026-06-10T16:48:44.105254+00:00",{"id":72,"slug":73,"title":74,"cover_image":75,"image_url":75,"created_at":76,"category":13},"741f86d7-bc7c-4ff6-8bc8-fbc0e7d780bd","gpu-programming-core-software-skill-en","GPU programming is becoming a core software skill","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1781109179855-erus.png","2026-06-10T16:32:18.834923+00:00",{"id":78,"slug":79,"title":80,"cover_image":81,"image_url":81,"created_at":82,"category":13},"2c661aae-7412-4250-a460-09791cc178ff","devin-pricing-turns-agents-into-seats-en","Devin’s pricing turns agents into seats","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1781096617614-fqw0.png","2026-06-10T13:03:11.058705+00:00",[84,89,94,99,104,109,114,119,124,129],{"id":85,"slug":86,"title":87,"created_at":88},"8008f1a9-7a00-4bad-88c9-3eedc9c6b4b1","surepath-ai-mcp-policy-controls-en","SurePath AI's New MCP Policy Controls Enhance AI Security","2026-03-26T01:26:52.222015+00:00",{"id":90,"slug":91,"title":92,"created_at":93},"27e39a8f-b65d-4f7b-a875-859e2b210156","mcp-standard-ai-tools-2026-en","MCP Standard in 2026: Integrating AI Tools","2026-03-26T01:27:43.127519+00:00",{"id":95,"slug":96,"title":97,"created_at":98},"165f9a19-c92d-46ba-b3f0-7125f662921d","rag-2026-transforming-enterprise-ai-en","How RAG in 2026 is Transforming Enterprise AI","2026-03-26T01:28:11.485236+00:00",{"id":100,"slug":101,"title":102,"created_at":103},"6a2a8e6e-b956-49d8-be12-cc47bdc132b2","mastering-ai-prompts-2026-guide-en","Mastering AI Prompts: A 2026 Guide for Developers","2026-03-26T01:29:07.835148+00:00",{"id":105,"slug":106,"title":107,"created_at":108},"3ab2c67e-4664-4c67-a013-687a2f605814","garry-tan-open-sources-claude-code-toolkit-en","Garry Tan Open-Sources a Claude Code Toolkit","2026-03-26T08:26:20.245934+00:00",{"id":110,"slug":111,"title":112,"created_at":113},"66a7cbf8-7e76-41d4-9bbf-eaca9761bf69","github-ai-projects-to-watch-in-2026-en","20 GitHub AI Projects to Watch in 2026","2026-03-26T08:28:09.752027+00:00",{"id":115,"slug":116,"title":117,"created_at":118},"9f332fda-eace-448a-a292-2283951eee71","practical-github-guide-learning-ml-2026-en","A Practical GitHub Guide to Learning ML in 2026","2026-03-27T01:16:50.125678+00:00",{"id":120,"slug":121,"title":122,"created_at":123},"1b1f637d-0f4d-42bd-974b-07b53829144d","aiml-2026-student-ai-ml-lab-repo-review-en","AIML-2026 Is a Bare-Bones Student Lab Repo","2026-03-27T01:21:51.661231+00:00",{"id":125,"slug":126,"title":127,"created_at":128},"6d1bf3f6-e191-4d30-b55b-8a0722fa6afe","ai-trending-github-repos-and-research-feeds-en","AI Trending Tracks Repos and Research Feeds","2026-03-27T01:31:35.709532+00:00",{"id":130,"slug":131,"title":132,"created_at":133},"010539a1-4c3a-4bd3-937a-26616422ee0d","awesome-ai-for-science-research-tools-map-en","Awesome AI for Science Is Becoming a Real Research Map","2026-03-27T01:46:50.89513+00:00"]