shubhamsaboo/awesome-llm-apps
20 skills63.8k installs2M starsGitHub
Install
npx skills add https://github.com/shubhamsaboo/awesome-llm-appsSkills in this repo
1Fullstack DeveloperFullstack-developer is an agent skill for solo and indie builders who need a coherent playbook across React/Next.js frontends, Node.js backends, and relational or document databases. It activates when you mention full-stack work, REST or GraphQL APIs, MongoDB, PostgreSQL, Express, or shipping a complete web product—not a single-file snippet. The skill encodes expert defaults: Composition-friendly React patterns, type-safe TypeScript on both tiers, schema validation with Zod, and authentication flows teams expect in SaaS. You use it while structuring features end to end: data models, API routes, UI consumption, and hooks for third-party integrations. It also nudges deployment and scaling considerations so the agent does not stop at localhost. It is instructional procedural knowledge, not a live MCP server; your agent still runs commands in your repo. Pair it with testing and security skills before production launch.6.5kinstalls2Academic Researcheracademic-researcher is an MIT-licensed agent skill from awesome-llm-apps that turns the agent into a cross-disciplinary research assistant. Solo builders use it whenever scholarly rigor beats vibe-based notes: conducting literature reviews, summarizing PDFs, critiquing methods, identifying gaps, and formatting citations in common styles. The SKILL.md defines a repeatable Paper Analysis Framework—question and significance, methodology with limitations, key findings, and field contribution—so outputs stay comparable across papers. Because triggers explicitly include research proposals and “when user mentions papers or scientific literature,” it is journey-wide rather than locked to ideation only; you can reuse the same structure in validate when scoping a feature against prior art, in build when writing technical background in docs, or in ship when documenting evaluation methodology. It does not fetch paywalled corpora by itself; you supply sources and the skill shapes analysis and prose.5.3kinstalls3Content CreatorContent Creator is an agent skill that turns vague topics into audience-focused writing for blogs, social networks, email, and marketing surfaces. Solo builders and indie marketers use it when they need blog posts, platform-specific social copy, newsletters, product descriptions, or headline work without outsourcing a writer. The embedded framework pushes you to define who you’re writing for, lead with a hook that promises delivered value, pack in actionable examples and takeaways, and format for skim reading with subheads and lists. It fits the growth and launch parts of the journey whenever distribution depends on credible, readable copy rather than code. Complexity stays approachable because the ritual is editorial, not technical—your agent applies consistent structure across channels while you supply product context and brand voice.4.2kinstalls4Technical Writertechnical-writer is an agent skill that applies professional documentation habits for solo builders shipping APIs, agents, and SaaS products who need READMEs, setup guides, API references, and tutorials without hiring a writer. It triggers when you mention documentation, need a changelog, or want complex ideas explained for developers and end users. The skill encodes four principles: lead with the reader’s goal, write plainly, show working examples with outputs and failure modes, and structure content from quick start to deep dives. That makes it useful during Build when you document architecture and endpoints, during Validate when you draft landing-adjacent explainers, and during Grow when you publish tutorials or release notes. It is a prose and structure workflow, not a doc-site deployer—pair it with your static site or repo docs folder.3.7kinstalls5Project PlannerProject Planner is an agent skill that acts as an expert project planner for solo and indie builders who need structure before implementation spirals. It applies when you are defining scope, creating a work breakdown structure, mapping dependencies, estimating timelines, or planning milestones and phases. The workflow walks from success criteria and constraints through deliverables, decomposes work into two-to-eight-hour tasks with testable completion, maps parallel work and critical path, then layers estimates and buffers. It fits SaaS, APIs, and agent-assisted products where one person must sequence backend, frontend, and launch work without a PM team. Use it during validation to lock scope and during build PM when reprioritizing sprints. The output is an actionable task graph—not a vague roadmap—so Claude Code, Cursor, or Codex can pick up ordered work with explicit dependencies rather than improvising from a one-line feature request.3.7kinstalls6Python ExpertPython Expert is a structured SKILL.md reference for AI agents that write or review Python. It ranks guidance by impact so agents fix correctness first—mutable defaults and error handling—then type safety with hints and dataclasses, then performance via comprehensions and context managers, and finally PEP 8 plus docstrings. Solo and indie builders shipping FastAPI scripts, data jobs, or agent tooling can invoke it whenever generated Python risks subtle bugs or inconsistent style. It is procedural knowledge, not an MCP or runtime tool: the agent applies the checklist inline during implementation or review. Pair it with ship-phase review skills when you want a separate diff report; use this skill when you want consistent Python craft embedded in every build session.3.6kinstalls7Code Reviewercode-reviewer is an agent skill that gives solo builders a prioritized review rubric instead of vague “look for bugs” prompts. It walks agents through CRITICAL security checks such as parameterized SQL and XSS-safe output, HIGH-impact performance issues like N+1 queries, HIGH correctness expectations around error handling, and MEDIUM maintainability habits including naming and typing. The structure mirrors how a small team would triage a pull request when nobody dedicated owns security or DBA review. Install it when you want consistent commentary on real vulnerability classes and data-access antipatterns before shipping SaaS, API, or CLI changes. It does not replace automated SAST or dependency scanning, but it steers conversational review toward fixes that materially reduce exploit and outage risk.3.4kinstalls8Data Analystdata-analyst packages expert-style guidance for SQL, Python pandas, and statistical analysis so a solo builder can explore messy datasets without hiring a dedicated analyst. Invoke it when you need extraction queries with JOINs and window functions, pandas pipelines for grouping and time series, or lightweight hypothesis and correlation work to support a product or growth decision. The skill emphasizes readable deliverables: commented SQL and code, example result shapes, performance caveats on heavy queries, and plain-language interpretation of what the numbers imply. It fits multiple journey moments—validating an idea with real usage data, building internal reporting in Ship-adjacent work, or compounding Grow metrics—but Prism shelves it under Grow analytics as the primary home for ongoing measurement and insight. It is procedural knowledge for agents, not a hosted notebook or warehouse connector; you still need database access and a Python environment where pandas runs.3.4kinstalls9Deep ResearchDeep Research is an agent skill that turns vague “look into X” requests into a disciplined investigation: clarify the question and required depth, map subtopics and open questions, then gather and weigh evidence from multiple credible sources before writing a synthesized summary with citations. It is aimed at solo and indie builders who use Claude Code, Cursor, or similar agents when scoping a product idea, sizing a market, comparing stacks, or backing launch and growth content with real references—not when they only need a one-line fact. The workflow emphasizes perspective diversity, publication currency, and explicit credibility judgment so downstream plans, landing copy, and specs rest on traceable reasoning. Use it whenever user language mentions research, investigation, in-depth analysis, or synthesized reporting with sources.3.2kinstalls10Ux DesignerUX Designer packages complete agent rules for solo builders who want product-quality interaction design without skipping research. Critical priorities start with user interviews and evidence-based personas, then WCAG AA accessibility and inclusive patterns before visual polish. The skill tells agents to ask how users last performed tasks, synthesize themes, and reject assumption-driven personas or contrast failures that block real users. It fits early validation when scoping flows and remains relevant while implementing frontend in Build and polishing Launch-facing screens. It is procedural knowledge—not a Figma plugin—so deliverables are research-aligned specs, accessibility checks, and design decisions agents can embed in tickets and components. Teams shipping fast still benefit because rules are explicit anti-patterns plus measurable targets like contrast ratios and minimum touch size.3.1kinstalls11Visualization ExpertVisualization Expert is a lightweight agent skill from awesome-llm-apps that acts as a chart-selection and design coach for solo builders presenting data. When you have query results, experiment outputs, or growth metrics but are unsure whether a pie chart will mislead or a scatter plot will clutter, the skill walks you through purpose-first choices: comparison bars, distribution histograms, relationship scatters, composition stacks, and trend lines. It encodes mainstream best practices—remove chart junk, avoid deceptive axes, and design for color-blind readers—so your dashboards in a SaaS admin or a launch blog post tell the truth simply. The agent responds with a recommended chart type, rationale, example code in common Python plotting libraries, and guidance on how to read the result. It fits indie operators who wear analyst and frontend hats and need fast, consistent visualization advice without hiring a data journalist. Use it while shaping Grow analytics narratives or when adding charts during Build frontend work.2.9kinstalls12Fact CheckerFact Checker is an agent skill that structures LLM-assisted verification so solo builders do not treat confident model text as ground truth. It walks through identifying the exact factual assertion, defining what evidence would prove or disprove it, gathering material from credible primary and authoritative sources, and issuing a rated conclusion with explicit confidence and reasoning. That discipline matters when you are researching a market myth, quoting a stat on a landing page, or sanity-checking a competitor’s viral claim before you repeat it in launch content. The skill is procedural rather than a live API integration—you bring URLs, documents, or searchable context, and the agent applies a consistent rubric instead of free-form debate. It pairs naturally with web fetching skills when you need source text first. Complexity is intermediate because good outcomes depend on source quality and transparent uncertainty rather than a single boolean answer.2.7kinstalls13DebuggerDebugger is an agent skill that teaches systematic root-cause analysis for software that misbehaves, crashes, or throws errors. It is aimed at solo and indie builders who ship with AI coding agents and need a repeatable ritual when something is “not working” instead of random edits. The skill applies when you are investigating bugs, reading stack traces, triaging logs, debugging performance regressions, or responding to production incidents—including intermittent failures that are hard to reproduce. You follow a structured flow: clarify expected versus actual behavior, collect error messages and environment facts, list hypotheses from most to least likely, then test them efficiently (including binary search across changes). That methodology fits multiple journey moments: failing tests before release, review-time regressions, and live error spikes after deploy. It does not replace your test suite or observability stack; it organizes how the agent thinks while you drive reproduction and verification.2.7kinstalls14Strategy AdvisorStrategy Advisor is an agent skill that turns vague business questions into disciplined strategic memos. Solo and indie builders use it whenever they need to compare paths—enter a niche, change pricing, pursue a partnership, or set a multi-quarter roadmap—without defaulting to gut feel. The skill walks through situational analysis (current state, stakeholders, market dynamics, competition, constraints), generates alternatives including unconventional ones, scores options against alignment, financial impact, resources, risk, and time horizon, then delivers a preferred recommendation with implementation notes and success metrics. It fits early journey research but remains useful at validate scope decisions and grow-phase pivots. Invoke it when the user mentions strategy, business planning, competitive analysis, trade-offs, or long-term planning rather than tactical implementation tasks.2.6kinstalls15Decision HelperDecision Helper is an agent skill that facilitates structured decision-making using five proven frameworks: simple pros/cons lists, weighted decision matrices, cost-benefit analysis, SWOT, and ICE scoring. Solo and indie builders install it when they face multiple viable paths—tech stack, pricing model, feature cuts, hiring vs DIY—and need comparable option write-ups instead of endless unstructured chat. The skill encodes when to apply each framework and a consistent markdown deliverable with named options, risks, effort levels, and matrix tables. It is methodology, not a data integration: your coding agent walks you through criteria, weights, and scores. Because invoke triggers span “making decisions” and “weighing trade-offs” without tying to one product phase, treat it as journey-wide shelf placement with Idea/research as the first natural stop. Pair with planning or brainstorming skills when a chosen option needs an implementation plan.2.5kinstalls16EditorEditor is a professional editing and proofreading skill that walks an agent through surface fixes, language polish, flow improvements, and structural revisions using a defined four-level model and multi-section checklist. Solo builders invoke it when landing pages, READMEs, newsletters, support macros, or launch posts need to read clearly and credibly without hiring a copy desk. It triggers on explicit edit, proofread, improve, revise, and readability language, and it separates proofreading from developmental work so the agent does not rewrite strategy when you only wanted comma fixes. Because written quality touches validation copy, build docs, ship release notes, and growth content, treat it as multi-phase: primary shelf in Grow content, with natural use when tightening validate landing text or launch distribution assets. Outcomes are tighter prose, consistent tone, and stronger structure—not fact sourcing unless copy editing explicitly calls for claim review.2.4kinstalls17Meeting Notesmeeting-notes is an agent skill that helps solo builders and tiny teams produce consistent meeting documentation without hiring an admin assistant. When you are taking notes, summarizing a discussion, or tracking follow-ups, the skill applies a single markdown structure: dated header, attendees, agenda, per-topic discussion, explicit decisions, a tabular action list with owners and deadlines, next steps, and a parking lot for deferred topics. It emphasizes actionable output—specific tasks, named owners, and timelines—rather than transcript dumps. The skill fits everyday agent chat: mention meeting notes or minutes and the agent shapes output for Slack, Notion, or repo docs. It does not replace a calendar or recorder; it formats and sharpens what you already captured during or after the call.2.4kinstalls18Sprint PlannerSprint Planner is an agent skill that acts as a scrum-style facilitator for indie and small teams who run two-week (or similar) delivery cycles without a dedicated PM. Invoke it when you need to turn a messy backlog into a sprint goal, a capacity-backed commitment, and a tabular backlog with owners and dependencies. It encodes Modified Fibonacci estimation, a standard capacity calculation, and velocity averaging so you do not over-commit when you are solo or part-time. The structured markdown output fits paste-in workflows for issues, docs, or chat handoffs to implementation agents. It is aimed at builders shipping SaaS, APIs, or agent features who already use agile language but want repeatable planning instead of ad-hoc chat. It does not replace your issue tracker; it produces the planning artifact you then execute in build and revisit during operate iterate cycles.2.2kinstalls19Email DrafterEmail Drafter is an agent skill that turns rough intent into structured professional email copy for solo and indie builders who wear sales, support, and founder hats. It encodes a repeatable framework—clear subject lines, appropriate greetings, concise bodies capped at two or three points, explicit calls to action, and professional sign-offs—plus tone guidelines so the same agent can shift from executive outreach to friendly team notes. Install it when you are staring at a blank compose window, need to request a meeting without sounding pushy, or must follow up after a demo or invoice without rambling. It is not a mail client integration; it produces draft text you paste into Gmail, Outlook, or your CRM. Because business mail shows up across validation conversations, launch outreach, and day-two customer support, the skill is tagged multi-phase with Grow lifecycle as the primary shelf. Pair it with your own facts and deadlines so numbers and dates stay accurate.2.1kinstalls20Content WriterContent Writer is an agent skill for solo builders who need marketing copy without hiring a copywriter. It targets landing pages, lifecycle emails, and social posts when you are creating promotional content, sales messaging, or brand voice for a SaaS or content product. The skill encodes a lightweight style guide: lead with the biggest benefit, prefer specific numbers over vague claims, end with a clear call-to-action, and avoid jargon unless your audience expects it. A explicit banned-word list filters AI-default hype, while headline and CTA templates give repeatable structure for launch and growth campaigns. Use it during validate when you need landing copy for a smoke test, at launch when distribution pages must convert, and in grow when you refresh emails or social hooks. It is not a full SEO technical audit or brand strategy workshop—it is fast, constraint-driven drafting your agent can iterate on in Claude Code, Cursor, or Codex.1.3kinstalls