davila7/claude-code-templates
111 skills86.6k installs3M starsGitHub
Install
npx skills add https://github.com/davila7/claude-code-templatesSkills in this repo
1Senior Data ScientistSenior Data Scientist is a methodology-oriented agent skill that packages world-class experiment design and feature engineering frameworks for indie builders shipping analytics or ML-backed products. It pushes production-first thinking from the first notebook: target roughly 10x headroom, 99.9% uptime goals, clear documentation, and full observability rather than throwaway prototypes. The skill walks through advanced patterns for distributed processing, low-latency real-time paths, and ML at scale with automation and monitoring hooks. Use it when you need a senior DS voice to structure experiments, engineer features responsibly, and align model work with deployable, secure systems—not when you only need a one-off chart. It complements agent workflows where you already have data and want disciplined design before implementation.2.7kinstalls2Senior BackendSenior Backend is an agent skill package aimed at solo and indie builders who need production-shaped APIs without hiring a staff engineer. It bundles procedural guidance with three Python utilities—an API scaffolder, a database migration analyzer, and an API load tester—so Claude Code or similar agents can generate project structure, harden Postgres usage, and validate throughput in one workflow. The documented stack spans NodeJS, Express, Go, Python, GraphQL, and REST, with emphasis on authentication, authorization, security implementation, and performance tuning. Use it when you are greenfielding a backend, refactoring slow queries, implementing domain logic, or asking an agent to review service-layer code against modern best practices. It complements frontend and DevOps skills by focusing on the persistence and contract layer that most indie products underestimate until launch week.2.2kinstalls3Excel AnalysisExcel Analysis is a compact agent skill that teaches pandas-first workflows for solo builders who live in spreadsheets for revenue tracking, cohort exports, or one-off operational dumps. Instead of asking the model to improvise brittle openpyxl snippets, the skill standardizes reading single or multi-sheet workbooks, computing groupbys and filters, adding derived columns like profit margin, and emitting new .xlsx files through ExcelWriter. It fits indie operators who validate pricing in validate with a model in Excel, analyze funnel exports in grow, or document recurring reports in build docs. The skill is procedural reference rather than a live MCP integration—you run Python in your environment. Intermediate complexity assumes comfort with DataFrames; beginners can still follow the copy-paste blocks. Pair it with your own visualization or BI stack when charts must leave the notebook.1.8kinstalls4Literature ReviewLiterature-review is an agent skill that supplies a long-form template for writing rigorous literature reviews in narrative or systematic styles. Solo builders and indie researchers use it when they need to document background, objectives, search methodology, and synthesized results in a format reviewers recognize, including optional PRISMA and PROSPERO fields. It helps agents turn scattered papers and notes into a single structured artifact with abstract, introduction, methodology, and conclusions sections rather than ad-hoc summaries. The skill fits idea-phase discovery when validating whether a niche is understudied or when preparing evidence for grant, thesis, or technical decision memos. It does not run database searches or perform citation management automatically; it standardizes the writing skeleton so human or agent effort focuses on content quality and inclusion criteria.1.6kinstalls5Senior FrontendSenior Frontend is a Claude Code Templates reference skill that equips solo builders and small teams with senior-level frontend discipline inside the agent loop. Rather than one-off snippets, it organizes repeatable patterns—each with description, applicable scenarios, TypeScript-oriented implementations, benefits, and trade-offs—plus cross-cutting guidelines on folder structure, naming, documentation, performance optimization, and client-side security. Common pattern catalogs and explicit anti-patterns help the agent avoid churny refactors and naive UI shortcuts. It is best invoked when you are actively building or refactoring React/Vue/Angular-style frontends and want opinions aligned with production maintenance, not tutorial-level demos. Complexity is advanced because the framing assumes you already ship UI and need consolidation, not hello-world teaching.1.6kinstalls6Market Research ReportsMarket Research Reports is a template-oriented agent skill that helps solo builders format market research writeups as professional LaTeX documents. It centers on the market_research.sty style package: a fixed color vocabulary for primary and secondary blues, greens for market metrics, oranges and reds for risk, and purples for recommendations, plus neutral grays for body hierarchy. Box environments such as key insight callouts give agents repeatable patterns instead of inventing layout from scratch each sprint. The skill fits early journey work when you are packaging competitor scans, TAM/SAM notes, or segment narratives for cofounders, advisors, or your own decision log. It assumes you already collected research content and need presentation discipline. It does not perform live market data collection or statistical modeling—that stays in research tooling elsewhere. Intermediate complexity reflects LaTeX fluency and comfort editing .sty-driven structures.1.4kinstalls7MarkitdownMarkItDown is an agent skill oriented around Microsoft’s MarkItDown library for converting common document formats into markdown text that LLM agents can read efficiently. Solo and indie builders use it when legacy content lives in PDFs, Word manuscripts, or PowerPoint decks and needs to become searchable context for coding, research synthesis, or documentation updates. The skill’s examples cover single-file conversion, saving outputs to disk, stream-based ingestion, and batch processing whole directories of research papers. That makes it especially useful during idea-phase literature review, build-phase doc migration, and grow-phase content repurposing without retyping sources. It is a procedural integration skill: your agent runs Python snippets or scripts you adapt from the examples rather than calling a separate MCP server. Expect intermediate comfort with Python paths and virtual environments; it does not replace dedicated OCR for scanned images unless your MarkItDown build supports those pipelines.1.4kinstalls8Exploratory Data AnalysisExploratory Data Analysis is an agent skill that walks a solo builder through inspecting a single dataset file and filling out a comprehensive markdown report. It standardizes filename, path, size, extension, and format metadata, then documents structure, dtypes, and dimensions before quality checks and descriptive statistics. The output mirrors what you would hand to a teammate or your future self before feature engineering, ML experiments, or product analytics. It is aimed at indie hackers and small teams who need repeatable EDA documentation without spinning up a notebook for every CSV or export. Use it when you receive an unfamiliar file, need to assess whether data is fit for downstream SQL or Python work, or want a citable summary of coverage and validity issues. The skill emphasizes transparency: placeholders make gaps obvious so you can fix ingestion or schema problems early in the build cycle.1.4kinstalls9Mobile DesignMobile Decision Trees is an agent skill that walks solo builders through structured choices for mobile apps: which framework fits OTA needs, branding, native hardware, and team skills. It frames React Native with Expo, Flutter, SwiftUI/UIKit, Jetpack Compose, and KMP as tradeoffs rather than defaults. Use it when an LLM might otherwise pick a trendy stack without matching constraints like app-store review cycles, pixel-perfect branding, or platform-specific APIs. The README stresses that outputs are reasoning aids for agents and humans planning architecture, navigation, and persistence—not finished project scaffolds. It pairs well with validate-phase scoping when you are still deciding native vs cross-platform, and with build-phase implementation when you need consistent answers across sessions.1.3kinstalls10Scientific WritingScientific-writing is an agent skill focused on citation styles for rigorous manuscripts—medicine, biomedical sciences, psychology, engineering, and chemistry each expect different in-text and bibliography rules. Solo builders writing literature reviews, grant appendices, ML evaluation papers, or health-tech documentation use it when they must not guess between Vancouver brackets and AMA superscripts. The guide opens with a comparative table linking AMA, Vancouver, APA, Chicago, IEEE, ACS, and NLM to disciplines and in-text formats, then drills into AMA with concrete examples for superscript placement relative to punctuation. It reinforces checking target journal author guidelines before locking a style—especially when PubMed-indexed journals prefer Vancouver or NLM. Complexity is beginner to intermediate: the rules are explicit, but edge cases (secondary citations, multiple works, online DOIs) still need journal-specific tweaks. Pair this skill with your reference manager export settings rather than treating it as automated bibliography generation.1.3kinstalls11Senior Data EngineerSenior-data-engineer is a pattern library skill aimed at solo and indie builders who outgrow ad-hoc SQL scripts and need enterprise-style data architecture without a full data platform team. It frames production-first modeling—scale headroom, reliability targets, security, and observability—then walks through distributed processing, real-time throughput, and operational ML concerns with monitoring and automation. Performance guidance emphasizes profiling, caching, and batching; reliability covers retries, circuit breakers, and continuous health checks. The content is intentionally domain-agnostic so agents can shape warehouse layers, streaming jobs, or analytics APIs during implementation, then reuse the same principles when hardening pipelines in production. Use it when your agent is sketching lakehouse tables, orchestrating ETL, or defending SLAs, not when you only need a single Postgres schema for a landing page MVP.1.3kinstalls12Stable Diffusion Image GenerationStable Diffusion Image Generation is an agent skill that walks solo builders through advanced Hugging Face diffusers workflows: loading Stable Diffusion v1.5 subfolders separately, assembling a StableDiffusionPipeline, and implementing a custom DDIM-style generate loop with guidance and dimensions. It targets indie makers who need branded visuals, app assets, or marketing imagery without outsourcing every render. Use it when your agent should write reproducible Python generation scripts rather than guessing API parameters. The guide emphasizes component-level control—scheduler choice, safety checker removal for controlled environments, and step-by-step denoising—so you can embed generation in CLIs, content pipelines, or internal tools. Expect intermediate-to-advanced ML familiarity; you are wiring models, not clicking a hosted UI.1.2kinstalls13Ui Ux Pro Maxui-ux-pro-max is a compact reference skill that maps common analytics questions to concrete visualization choices for solo builders shipping dashboards, admin panels, or growth metrics views. Each row ties keywords like trend, compare, part-to-whole, correlation, and heatmap intensity to a best-fit chart type, fallback options, palette guidance, performance expectations, accessibility caveats, recommended JavaScript chart libraries, and suggested interaction depth from hover-only through zoom and brush. Agents consult it when a founder asks for “a chart for signups over time” or “how to show plan mix without an unreadable pie,” avoiding generic defaults that fail colorblind users or melt down on large scatter datasets. It does not replace a full design system; it accelerates the decision layer before you wire Recharts or Chart.js into your Next or React app, keeping choices explicit enough to cite in PR descriptions or investor-facing product demos.1.2kinstalls14Ui Design Systemui-design-system packages a procedural design-token generator for solo and indie builders who need a credible visual foundation without hiring a design ops team. You supply a brand hex (default #0066CC) and a style label; the skill walks an agent through producing a full token object covering color ramps, typography rhythm, spacing on an eight-point grid, borders, elevation shadows, motion curves, responsive breakpoints, and z-index layering. That artifact becomes the single source of truth when you scaffold Tailwind themes, CSS custom properties, or hand off to a component library in Claude Code, Cursor, or Codex. Use it early in the build phase whenever screens feel ad hoc, when rebranding a landing page, or when two features diverge on spacing and color. It is a generator skill, not a critique or accessibility audit—pair it with review skills before ship if you need WCAG validation.1.2kinstalls15WhisperWhisper Language Support is a reference skill for solo builders adding speech-to-text to agents, note apps, podcasts, or support tools. It summarizes OpenAI Whisper’s multilingual coverage so you can pick default languages, set user-facing locale lists, and set realistic quality expectations before you commit to a single-model deployment. Top-tier languages are flagged where word error rates are typically strongest, while a second tier signals markets that may need human review or domain-specific fine-tuning. The enumerated ninety-nine language list helps agents generate consistent ISO codes in config, API payloads, and analytics—not ad-hoc country names. Use it when scoping an MVP voice feature, comparing markets for a validate prototype, or documenting which locales your grow-stage content pipeline will auto-transcribe. It does not replace Whisper installation or GPU tuning; it gives your coding agent accurate language facts to embed in integration code and product copy.1.1kinstalls16Scientific VisualizationScientific-visualization is an agent skill template for solo builders and indie researchers who ship figures with Python and matplotlib. It encodes widely cited colorblind-friendly palettes—Okabe-Ito, Wong, and multiple Paul Tol sets—plus named sequential colormaps such as viridis so agents do not default to unreadable rainbow schemes. The workflow centers on importing palette dictionaries or lists and calling apply_palette before plotting lines, bars, or heatmaps. Use it whenever an agent drafts publication plots, notebook EDA charts, or dashboard exports where accessibility and print legibility matter. It complements generic styling skills by focusing on evidence-based color choices for continuous, diverging, and categorical scientific data rather than brand aesthetics alone.1.1kinstalls17Scientific SlidesScientific-slides is an agent skill from the davila7 Claude Code templates collection aimed at solo builders, indie researchers, and technical founders who must produce credible scientific or evidence-heavy presentations quickly. Install it when you already have findings, specs, or draft prose and need slides that follow academic conventions rather than startup pitch clichés. The skill packages procedural guidance for structuring slides around hypotheses, methodology, data, limitations, and takeaways so your coding agent does not improvise vague bullet lists. It sits in the Build phase under documentation because the primary output is a communicable artifact you reuse in validate pitches, launch webinars, and grow content—but the skill itself is task-focused on deck generation. Pair it with your repo notes or paper drafts; review figures and citations yourself before presenting. Complexity is beginner-friendly if you supply clear source material; confidence is moderate because public SKILL.md detail in the catalog is thin compared to name and install footprint on skills.sh.1kinstalls18Senior DevopsSenior DevOps is an agent skill that packages a CI/CD pipeline reference guide for solo and indie builders who own the full stack from commit to production. It walks through repeatable patterns—when to apply them, example implementations, benefits, and trade-offs—alongside organization, performance, and security guidelines. Use it when you are designing or refactoring pipelines, hardening deploy paths, or teaching your coding agent consistent DevOps vocabulary instead of one-off chat advice. The content is template-shaped (patterns A–C, anti-patterns) rather than a single vendor integration, so it fits SaaS, API, and CLI products equally. It complements testing and review skills by focusing on how code leaves your machine reliably, with monitoring and scaling called out for after launch.1kinstalls19Citation ManagementCitation-management is a Claude Code Templates skill that equips solo builders with a BibTeX-oriented reference file showing properly formatted scholarly entries. When you document APIs, write research summaries, or ship content that must cite primary literature, inconsistent citation metadata undermines trust and publication workflows. This skill centers on exemplar @article records—with author lists, journal titles, volume, issue, pages, and DOI—so your agent can mirror the same structure for new sources instead of inventing ad hoc formats. It fits builders who export to LaTeX, Pandoc, or static-site bibliographies rather than teams running a full reference manager UI. Use it while drafting technical documentation or long-form posts where every claim should trace to a citable source. The material is template-first: extend the patterns for books, proceedings, and web sources as your doc set grows.1kinstalls20Marketing Strategy PmmMarketing Strategy PMM is an agent skill that walks solo and indie builders through product marketing fundamentals: ideal customer profiles, differentiated positioning, go-to-market planning, competitive analysis, and launch messaging. It encodes frameworks such as April Dunford positioning alongside practical artifacts like battlecards, launch playbooks, and market-entry notes. Use it when you are still proving an idea, reframing a feature set, or preparing a first real launch—not when you only need a single ad headline. The skill fits builders shipping SaaS or content products who wear the PMM hat without a dedicated marketing team. It connects research and validation work to distribution and lifecycle messaging so agents produce structured strategy copy instead of generic marketing fluff.1kinstalls21Ux Researcher DesignerUX Researcher Designer packages a data-driven persona generator for solo builders who have some user interviews, survey snippets, or behavioral tallies but no formal research team. The embedded Python PersonaGenerator class defines persona components—demographics, psychographics, behaviors, and needs—and applies archetype templates such as power user and casual user with characteristic lists, goals, frustrations, and representative quotes. You feed structured or semi-structured research inputs; the skill guides synthesis into personas you can cite in validate-phase scope conversations and build-phase UX work. It suits indie SaaS and content products where one founder must stand in for researcher and designer. Complexity is intermediate because you need readable Python, clean input data, and judgment when mapping clusters to archetypes. Use early in the journey when positioning is fuzzy; pair outputs with landing copy and prototype task flows rather than treating personas as permanent truth without fresh data.963installs22RemotionThis Remotion 3D skill packages procedural knowledge for embedding Three.js scenes in Remotion using React Three Fiber and @remotion/three. Solo builders shipping programmatic video—explainers, social clips, or SaaS-generated previews—install it so agents enforce hard rules that generic Three.js tutorials miss: wrap scenes in ThreeCanvas with video dimensions, light meshes properly, and drive every motion from useCurrentFrame instead of autonomous useFrame loops that break deterministic renders. The readme documents package installation across package managers and shows a minimal red sphere example. Complexity sits at intermediate: you need an existing Remotion project and comfort with React video composition. Use it while coding Build-phase frontends, then reuse the same constraints when you iterate Launch creatives or Grow content batches. It is not a full modeling or asset pipeline guide; it is the guardrail layer between R3F habits and Remotion’s frame-based renderer.941installs23Mermaid Diagram SpecialistMermaid Diagram Specialist is an agent skill that turns plain-language system and process descriptions into production-ready Mermaid syntax embedded in markdown. Solo builders use it when architecture, APIs, or user flows are easier to ship as diagrams than walls of prose—during README work, internal ADRs, or stakeholder-facing technical pages. The skill enumerates when to invoke it from its overview: architecture documentation, workflow visualization, ERDs, sequence flows, state machines, component relationships, decision trees, and user journeys. It states prerequisites such as a clear spec to document and the target diagram type, and outputs typed Mermaid blocks suitable for GitHub, Notion, or static doc sites. It fits indie teams who want consistent visual docs without opening a separate design tool for every change.926installs24Prompt EngineerPrompt Engineer is an agent skill for solo and indie builders who treat prompts like code: clear roles, constraints, and measurable behavior. It walks you through structured system prompts, few-shot sets, chain-of-thought when appropriate, and evaluation loops so small wording changes do not silently break agents in Claude Code, Cursor, or Codex workflows. Use it when you are shipping chat features, coding assistants, or internal tools that need stable JSON, tool-use discipline, or repeatable tone. The skill emphasizes iteration with explicit success criteria, token-aware context packing, and negative constraints so models stop doing the wrong thing. It fits the Build phase first but also applies while validating prototypes and reviewing agent outputs before ship.911installs25Generate Imagegenerate-image is a Python agent skill that wraps OpenRouter’s image generation and editing APIs for solo builders who need illustrations, mocks, or refreshed creative without maintaining bespoke HTTP glue. You point the script at a text prompt, optionally attach a source image for edits, and select from models such as Gemini image preview and Flux variants listed in the skill readme. API keys are resolved from OPENROUTER_API_KEY in a nearby .env file, which keeps secrets out of chat transcripts when your agent runs shell commands. Use it during Build when you need integration-backed assets, during Validate when prototyping landing visuals, or during Launch and Grow when refreshing thumbnails and social creatives. It is a task-oriented generator, not a full design system—pair it with your brand guidelines in the prompt for consistent results.902installs26Senior QaSenior Qa is an agent skill packaged as a testing toolkit for solo and indie builders shipping JavaScript full-stack apps. It targets the moment you need a credible QA plan—not a single happy-path test—covering strategy design, test case writing, automation implementation, manual passes, and coverage review for React, Next.js, and Node environments. The skill routes work through three scripted capabilities: generating structured test suites from a project path, analyzing coverage and quality metrics on a target tree, and scaffolding E2E harnesses when browser or integration flows matter. Use it while you are still implementing features (to add tests alongside code) and again before you tag a release (to close coverage gaps and document regression paths). It does not replace a dedicated CI platform or production observability; it accelerates the artifacts and analysis that make ship-phase testing repeatable for agent-assisted development.895installs27JiraJira is an agent skill that lets solo builders manage Atlassian Jira from chat using either the jira CLI or Atlassian MCP tools. On each run it detects which backend is available and routes commands or tool calls accordingly; if neither exists, it walks through setup instead of failing silently. It covers viewing and listing issues, creating work items, transitioning states, assigning tickets, and checking sprint-oriented lists using patterns like issue keys and keywords such as ticket, sprint, and backlog. For indie teams shipping features while wearing the PM hat, it reduces context switching between the terminal, browser, and the agent. It is an integration-style skill meant for day-to-day ticket hygiene during build and operate cycles, not a replacement for Jira admin or workflow design.887installs28Video DownloaderVideo Downloader is a lightweight agent skill from Claude Code Templates that lets a solo builder pull videos from YouTube and other supported platforms to local disk with explicit quality and format choices. It fits growth and content workflows—archiving webinars, saving course material, grabbing b-roll for edits, or backing up your own uploads—while supporting batch playlists and metadata retention. The skill is procedural: you describe the URL, desired resolution, or audio-only export and the agent orchestrates the download stack rather than exposing a single fixed binary. It can also support Build-phase integration work when you are wiring media ingestion into a content pipeline. Respect platform terms of service, copyright, and consent before downloading third-party material; Prism treats this as automation for content ops, not a bypass for DRM or licensing rules.853installs29Senior ArchitectSenior-architect is a reference-oriented agent skill from the Claude Code templates line that packages architecture patterns and practices for solo and indie builders who need production-grade structure without a dedicated staff architect. The SKILL.md-style body walks through named patterns with descriptions, implementation sketches, benefits, and trade-offs, plus cross-cutting guidelines on code organization, performance, and security. It also calls out common patterns to reuse and anti-patterns to avoid, with a short tools and resources section for follow-up research. Invoke it when you are greenfielding a service, refactoring a monolith, or preparing a feature that will affect auth, data boundaries, or scaling. It is editorial reference material rather than a one-click generator: you still make final calls on cloud vendor and team constraints. Pair it with planning and review skills when you want the agent to propose options, then sanity-check the chosen design against these guardrails before implementation.833installs30Scientific Critical ThinkingScientific Critical Thinking is a knowledge-oriented agent skill from the Claude Code templates collection that encodes how cognitive and incentive biases distort scientific work—from confirmation bias and HARKing to publication bias and grey-literature gaps. Solo builders shipping AI research tools, health analytics, or any product that cites studies need a disciplined lens before they treat LLM summaries as evidence. The skill is journey-wide: you can invoke it while exploring an idea, scoping a validate prototype, reviewing ship-ready claims, or iterating on data features in production. It does not fetch papers or run stats; it trains the agent to challenge study design, preregistration, null-result visibility, and interpretation hygiene. Pair it with your own literature tools when you need citations, but use this skill whenever conclusions outrun the underlying methods.829installs31Senior Computer VisionSenior Computer Vision is an agent skill that encodes world-class CV architecture judgment for solo builders and small teams shipping detection, segmentation, or embedding workloads beyond notebook prototypes. It steers conversations toward production realities—how inference scales under 10x load, how real-time paths stay within latency budgets, and how batch pipelines tolerate partial failures. The material spans distributed data processing, streaming or edge-adjacent serving, and operational ML habits such as continuous monitoring and automated retraining hooks. Security and privacy are treated as first-class: validated inputs, controlled access to imagery, and audit trails for regulated or customer data. For indie SaaS or internal tools that add vision features, the skill complements generic coding skills by naming tradeoffs explicit seniors make—when to cache embeddings, when to fuse models, and how to document contracts for downstream API consumers. Use it while designing services in Build, hardening paths in Ship, and tuning cost or SLOs in Operate.810installs32MuiThe MUI skill packages institutional knowledge for Material-UI v7 in React: when to reach for the sx prop, how to compose core components, and how to wire themes and responsive breakpoints without fighting v7 breaking changes. Solo builders shipping dashboards, internal tools, or customer-facing SaaS use it whenever UI work touches MUI—not generic CSS frameworks—so agents suggest patterns that match library conventions rather than one-off inline hacks. It spans styling decisions, layout with Grid and Box, dialog and card patterns, and hooks like useTheme and useMediaQuery for adaptive interfaces. Reach for it during active frontend implementation or a controlled v6→v7 migration; skip it for non-React stacks or design systems outside MUI. The skill is reference-heavy and task-oriented: identify the MUI task, then apply the matching pattern from the skill body.808installs33Google AnalyticsGoogle Analytics is an agent skill package for wiring Google Analytics 4 to the Analytics Data API and executing structured performance reviews. Solo and indie builders who ship web products but lack a dedicated growth analyst use it to stand up service-account auth, load property configuration safely, and follow repeatable analysis playbooks—such as a 30-day traffic overview with prior-period comparison—instead of improvising API calls in chat. The skill emphasizes credential hygiene (JSON keys and .env never in version control) and numeric GA4 property IDs versus measurement IDs. It fits the Grow phase when you need defensible numbers for roadmap calls, SEO experiments, or lifecycle email decisions. Intermediate complexity: you must complete Google Cloud IAM, API enablement, and GA4 property access grants before the agent can query data reliably.798installs34Ceo Advisorceo-advisor is a lightweight Claude Code template skill packaged in davila7/claude-code-templates. Prism lists it for solo and indie builders who want an agent persona oriented toward executive-level questions: market focus, roadmap bets, hiring versus automation, pricing posture, and what to cut before a build sprint. The published SKILL stub is metadata-heavy, so treat the skill as a starting persona and prompt scaffold you customize in your repo rather than a fully specified multi-step ritual. Invoke it when you are validating scope or replanning growth and operations strategy and want structured pushback instead of generic coding assistance. Combine with your own context documents (metrics, runway, competitive notes) for usable output.796installs35Senior Fullstacksenior-fullstack is a Claude Code Templates reference skill that encodes architecture patterns, implementation examples, guidelines, and anti-patterns for experienced fullstack work. Solo builders wearing every hat use it when an agent needs a structured lens for code organization, performance bottlenecks, scaling approaches, and security basics—not a single CRUD tutorial. The content spans pattern descriptions with benefits and trade-offs, TypeScript-flavored samples, and explicit what-not-to-do sections. It is most often invoked during build when shaping modules and APIs, but the performance and security sections naturally support ship and operate conversations. Treat it as procedural knowledge to steer agent output toward maintainable fullstack designs rather than as an automated deploy or test runner.793installs36Matplotlibmatplotlib is a compact agent skill that mirrors a matplotlib quick-reference: how to create Figures and Axes, call core plotting methods, tune aesthetics, and export publication-ready images. Solo builders using Claude Code or Cursor to ship analytics scripts, internal dashboards, or ML evaluation plots can invoke it whenever the model needs correct method names, kwargs, or layout patterns without hallucinating deprecated pyplot flows. The skill stays in the Build phase as documentation-backed agent tooling, but the same reference supports Grow analytics storytelling and Validate prototype charts. It does not run code or install packages; it standardizes how agents structure plots so your repo gets consistent, idiomatic matplotlib instead of one-off pseudo-API snippets.786installs37Scientific Schematicsscientific-schematics is an agent skill from Claude Code Templates that wires solo builders into an AI-assisted pipeline for scientific figures. Instead of hand-drawing CONSORT flows, pathway diagrams, or study schematics in a desktop tool, you describe the diagram in prose and run the bundled Python generator against OpenRouter. The included bash example walks through API key checks, figures/ directory creation, and a CONSORT flow with numbered cohorts, exclusions, and final analysis arms—matching how indie researchers and health-tech founders document trials and methods. Iteration flags produce multiple PNG versions and a JSON review log so you can pick the best render before dropping assets into a paper, investor appendix, or product validation doc. Expect intermediate setup: environment variables, pip dependencies, and paid model access. Use it when text-to-figure speed beats manual vector editing; skip when you need regulated figure tooling with locked templates and audit trails inside a certified authoring suite.780installs38Seo OptimizerSEO Optimizer is a comprehensive agent skill that acts as a search specialist for solo builders who own their own marketing stack. It consolidates keyword strategy, on-page optimization, technical SEO, structured data, and performance considerations into one procedural reference your coding agent can follow while you ship sites, blogs, or SaaS marketing pages. Use it when you are improving rankings, researching long-tail opportunities, fixing meta and schema gaps, or auditing an existing property—activities that start in Launch but recur in Grow as you publish content and in Ship when page speed and Core Web Vitals block indexing. The skill does not replace dedicated rank trackers or automated crawlers; it tells the agent how to think and implement SEO best practices consistently. Ideal when organic search is a primary acquisition channel and you need repeatable checklists instead of one-off chat advice.779installs39Senior Prompt Engineersenior-prompt-engineer packages senior-level guidance for designing agentic systems and prompt engineering workflows when you are past toy demos and need something an indie team can run in production. The material stresses production-first defaults—plan for higher load, design for failure with retries and circuit breakers, and instrument systems so you can see regressions before users do. It outlines reusable architecture patterns for distributed and real-time workloads and connects prompt design to broader ML-at-scale concerns such as monitoring and automation. A distinct thread covers LLM evaluation frameworks so you can judge whether prompt or agent changes actually improve outcomes instead of guessing from one chat transcript. It reads like a condensed playbook rather than a step-by-step checklist, which suits builders who already ship agents but want a structured rubric when refactoring prompts, adding tools, or hardening an agent service. Expect to adapt the patterns to your stack; the skill does not replace domain-specific integration skills.764installs40Crewai Multi Agentcrewai-multi-agent is a procedural guide for building event-driven multi-agent pipelines with CrewAI Flows. Solo and indie builders shipping agent products use it when sequential Crews are not enough—conditional paths, durable flow state, parallel start points, and listeners that react to prior steps. The skill walks through defining a Pydantic state model, subclassing Flow with typed generics, wiring @start and @listen methods, and running kickoff() while inspecting final state. It contrasts Crews (simple collaboration, hierarchy) with Flows (branching, complex state, event-driven and hybrid designs). Expect Python examples aligned with crewai.flow imports rather than one-off prompt tricks. Best paired with an existing CrewAI or Python agent repo where you are ready to refactor orchestration into explicit graphs instead of ad-hoc agent chatter.763installs41Clean CodeClean-code is a journey-wide agent skill that installs pragmatic engineering standards for solo builders who ship with Claude Code, Cursor, Codex, and similar tools. It compresses classic craft into actionable tables: single responsibility, no duplication, simplest working design, no speculative features, and leaving files slightly better than you found them. Naming and function-size rules target the most common agent failure modes—cryptic identifiers, mega-functions, and comment-as-documentation—so refactors stay cheap. Because it is framed as CRITICAL procedural knowledge rather than a linter, you invoke it whenever the agent writes or edits production code, during implementation spikes, and again when you review diffs before release. It pairs naturally with dedicated review or security skills but does not replace them; it shapes how every change is written so later audit passes are easier.750installs42Quality Documentation ManagerQuality Documentation Manager is a template-oriented agent skill from Claude Code Templates that helps solo builders organize documentation programs: what lives in SKILL.md versus on-disk assets, and where long-form reference material should sit. The bundled readme is largely scaffold text—meant to be replaced with your real brand guidelines, API references, or status-update playbooks—but it encodes useful conventions from other skills (product-management references, BigQuery examples). Use it when you are bootstrapping a skill repo or product docs folder and need a consistent split between procedural instructions and files the agent outputs without loading megabytes into context. It suits indie teams formalizing doc quality before ship reviews, not as a finished linter or doc generator. Expect to substitute placeholders with your templates and delete unused asset sections.748installs43Senior Securitysenior-security is a procedural reference for cryptography implementation and broader application-security engineering, packaged for agents helping solo builders ship SaaS, APIs, and CLIs without reinventing secure defaults. Invoke it when you are wiring encryption, hardening auth flows, structuring security modules, or reviewing code for common vulnerability classes. The skill organizes guidance into implementable patterns with TypeScript examples, plus explicit anti-patterns so the agent avoids risky shortcuts. It spans Build when you encode crypto and access control, Ship when you run security-focused implementation and review, and Operate when you maintain secrets and data-protection behavior in production. Complexity is advanced: you should already understand threat basics and language-level crypto APIs. It is a knowledge skill, not an automated scanner—pair it with tests and real security review for high-risk systems.740installs44Agile Product OwnerAgile-product-owner is a template-oriented agent skill backed by a User Story Generator that drafts INVEST-aligned stories from defined personas and story patterns. Indie founders who wear the PM hat use it when a feature idea needs crisp “As a … I want … so that …” wording plus persona-specific needs—efficiency for end users, control for admins, guidance for new users—before coding agents touch the repo. The embedded Python class models feature, improvement, and fix templates so outputs stay comparable across epics. It fits early validation when you are narrowing MVP scope and again during build when you split work for implementation plans or PRs. It does not replace discovery interviews or analytics; it accelerates consistent backlog artifacts. Intermediate complexity if you extend personas or wire the generator into your own tooling.739installs45Langchainlangchain is a procedural agent-building skill for solo developers who want a concise LangChain map instead of piecing together scattered docs. It explains what agents are in the ReAct sense— the model plans, calls tools, reads results, and loops—and walks through creating an agent with create_agent, attaching Python-callable tools, and running invoke with a message list. Examples highlight Anthropic Claude Sonnet and OpenAI GPT-4o as reasoning engines, with notes on temperature and when stronger models help complex reasoning. Use it in the Build phase while you are implementing chatbots, internal copilots, or CLI assistants that must combine LLM judgment with calculators, search, or your own functions. The skill is instructional reference, not a deployed runtime; you still own dependency versions, tool safety (especially eval-based demos), and production observability. Pair it with your project’s testing and secrets practices before shipping agent endpoints to users.738installs46Qdrant Vector SearchQdrant Vector Search is an agent skill from Claude Code templates that teaches solo builders how to run Qdrant beyond a single container: multi-node Docker Compose with Raft consensus, bootstrap chains, and Python client calls for collections with sharding options. It targets indie AI products that need reliable semantic retrieval for agents, RAG pipelines, or recommendation features without guessing cluster env vars. You follow explicit YAML for three nodes, storage volumes, and port layouts, then apply client code patterns for VectorParams and sharding. The skill fits builders already embedding embeddings who must graduate from localhost Qdrant to a small HA cluster before ship. It is integration-heavy—not a hosted SaaS substitute—so you still own networking, backups, and security on your infra.729installs47Conversation MemoryConversation Memory is an agent skill for solo builders shipping assistants that must feel continuous across sessions, not amnesiac on every thread. It treats memory as retrieval and relevance engineering—not a dump of chat logs—by guiding short-term working context, durable long-term recall, and entity-centric fact stores that can be updated without duplicating noise. You get concrete patterns for tiered systems, entity updates, and prompt assembly so the model sees what matters now without the failure modes the skill documents: unbounded stores that slow everything down, or bulk recall that drowns the user’s actual question. Use it when wiring persistence for Claude Code, Cursor, or Codex workflows where users expect the agent to remember preferences, prior decisions, and account facts across days or months, and you need guardrails on when to remember, consolidate, and forget.725installs48Scientific Brainstormingscientific-brainstorming equips solo builders and researcher-founders with advanced, named brainstorming frameworks tuned for scientific work. The SKILL body positions it as a reference when everyday ideation stalls or when someone explicitly wants a methodology like SCAMPER. Each SCAMPER letter comes with generic prompts plus scientific applications—replacing catalysts, merging omics layers, borrowing algorithms from other fields—so your agent does not produce vague “think outside the box” lists. Because ideation quality gates everything downstream, the skill is journey-wide: you can reopen it before validating scope, before choosing an MVP instrument, or when iterating on experiments in Operate. It pairs well with generic brainstorming or writing-plans skills when the output must become a testable plan. For Prism users shipping AI-assisted lab tools, bioinformatics utilities, or research agents, this skill keeps exploration rigorous enough to cite in notes and proposals without forcing premature implementation.722installs49Xlsxxlsx is an agent skill from the Claude Code templates line for working with Excel (.xlsx) workbooks programmatically. Solo and indie builders install it when spreadsheets are still the contract with finance, ops, or partners—forecast models, export tabs, bulk updates, or ingesting uploaded sheets—while the agent keeps changes in the repository context. The skill embeds procedural knowledge for spreadsheet tasks so Claude Code, Cursor, or similar agents do not improvise fragile CSV hacks when the deliverable must open correctly in Excel. It is most natural during Build when you integrate exports or validate structured inputs, and it resurfaces in Grow or Operate when you refresh metrics workbooks or incident runbooks stored as files. Complexity assumes comfort running Python helpers the skill documents and reviewing cell-level changes before commit. Treat licensing restrictions seriously: Anthropic’s skill terms limit extraction and redistribution outside authorized services. Prism lists it as a productivity integration, not a hosted spreadsheet API.707installs50Pptxpptx is an agent skill that teaches a template-driven workflow for editing Microsoft PowerPoint files with Python helpers rather than guessing layouts in prose. Solo builders use it when they have an existing .pptx brand deck and need new content—pitch, launch recap, webinar, or sales one-pager—while preserving master layouts and visual variety. The skill stresses analyzing thumbnails and markitdown text first, mapping each content section to a distinct layout (multi-column, full-bleed image, quote, stat callout), then unpacking the OOXML, completing structural slide changes before touching copy. That ordering prevents broken decks from half-finished reordering. Where subagents exist, parallel per-slide XML edits speed large decks. Expect shell access to run scripts under scripts/. The skill fits validate pitch drafts, launch announcement decks, and grow/content repurposing of talks into slide format.705installs51Frontend DesignFrontend Design is an agent skill from the Claude Code Templates collection for solo builders who want their shipped UI to feel intentional rather than interchangeable. Prism lists only sparse README text for this package, so treat it as a design-oriented template skill you invoke while you are in active frontend Build work: before or during layout passes, component styling, and marketing-page polish. Pair it with your stack-specific frameworks (React, Vue, plain HTML) by asking the agent to honor the skill’s design rules while generating markup and styles. It is not a component library install and not automated visual regression testing—it is procedural guidance the agent follows so typography, spacing, and aesthetic choices stay coherent. Re-read the full SKILL.md in the repo before production use because the public ingest did not include the detailed checklist body.698installs52Code ReviewerCode Reviewer packages a pragmatic review pipeline for solo builders who cannot afford a standing senior reviewer on every PR. The skill centers on three runnable Python scripts: one to analyze pull requests, one to deep-check code quality with metrics and recommendations, and one to generate review reports you can attach to GitHub or share with your agent. It targets TypeScript, JavaScript, Python, Swift, Kotlin, and Go, so indie teams shipping a polyglot stack still get one repeatable ritual. Invoke it when you are reviewing merges, hunting regressions before release, or standardizing feedback language across contributors. It is not a substitute for domain expertise on business logic, but it accelerates the tedious pass—style, obvious bugs, security hints, and checklist coverage—so you spend review time on product decisions instead of re-typing the same comments.697installs53Scikit LearnScikit-learn is a reference-oriented agent skill that walks solo builders through model selection and evaluation using scikit-learn's model_selection module. It covers basic and stratified train_test_split, three-way splits for honest validation, and cross-validation with KFold and StratifiedKFold—including when stratification matters for imbalanced classification. The skill is aimed at indie developers who are shipping ML-powered features in a SaaS, API, or internal tool and need procedural guardrails instead of guessing split ratios or leaking test data. Use it while prototyping in validate and while hardening pipelines in build so your agent proposes sklearn-idiomatic code with fixed seeds and explicit fold semantics. It does not replace domain feature engineering or production monitoring, but it shortens the loop from raw arrays to defensible offline metrics before you commit to a model in production.692installs54Draw Iodraw-io is an agent skill for producing and maintaining draw.io diagrams without hand-drawing every box in a GUI solo. It treats .drawio files as structured XML so an agent can reposition elements, enforce margins, and keep typography consistent—especially when you embed diagrams in Quarto decks or long-form docs. You reach for it when you need architecture overviews, flowcharts, or AWS-style reference diagrams that must stay version-controlled and re-exportable to PNG for slides, Notion, or static sites. The workflow covers programmatic layout, icon integration, and conversion scripts or pre-commit hooks so PNGs do not drift from source. For indie builders shipping APIs, agents, or small SaaS stacks, it turns “diagram debt” into a repeatable doc artifact instead of a one-off screenshot scramble.674installs55Senior Ml EngineerSenior-ml-engineer is a reference and workflow skill that packages world-class guidance for senior ML and AI engineers integrating LLMs and running models in production. Solo builders shipping inference APIs, agent backends, or data products use it when they need more than notebook code—fault-tolerant distributed jobs, real-time serving constraints, and MLOps habits like continuous monitoring and automated retraining hooks. The material stresses designing for tenfold load headroom, profiling before optimizing, strategic caching, and security-by-default on prompts and payloads. It is not a single CLI integration; it steers agent conversations toward repeatable architecture decisions across implementation, hardening, and live operations. Pair it with concrete repo skills for training code or cloud deploy, and treat this skill as the senior lens on patterns, tradeoffs, and operational readiness before committing to stack choices.669installs56Docker ExpertDocker Expert is an agent skill that acts as a hands-on containerization specialist for solo and indie builders shipping with Claude Code, Cursor, or similar agents. It is meant to be invoked proactively when Dockerfiles are bloated, builds fail mysteriously, Compose wiring is wrong, or you need production-grade security and networking patterns before launch. The skill prioritizes fast repo analysis with read and search tools, then targeted shell diagnostics such as version checks and image inspection, rather than generic advice. It covers multi-stage builds, layer caching discipline, minimal base images, non-root users, secret handling, health checks, and Compose service design aligned with current practice. When the problem is outside pure Docker—Kubernetes control planes, CI pipelines in GitHub Actions, managed ECS/Fargate, or persistence-heavy database containers—it tells the agent to switch experts instead of guessing. That scope boundary keeps answers actionable for one-person teams who still need reliable containers without running a platform team.665installs57PdfPdf is a journey-wide agent skill from the Claude code templates collection for working with PDF documents inside agent workflows. Solo builders encounter PDFs everywhere—signed contracts, RFC exports, academic papers, scanned invoices, and customer SOWs—and this skill standardizes how the agent should approach extraction, summarization, and question answering over those files. It fits whenever prose is locked in a non-editable format during Idea research, Validate scoping, Build implementation, or Grow content repurposing. The package carries Anthropic’s proprietary license notice: usage is governed by your Anthropic agreement and includes restrictions on redistribution and derivative extraction outside authorized services. Treat it as procedural guidance plus compatible tool invocation patterns rather than a standalone PDF library. Complexity is beginner for “summarize this PDF” tasks but rises when you chain PDF output into code or structured data. Always review the Security Audits panel before installing third-party template forks.658installs58Marketing IdeasMarketing-ideas is an agent skill that acts as a strategist with a fixed library of 140 marketing approaches tailored to SaaS and software products. Solo builders invoke it when they have something to ship but lack a marketing playbook—whether at first launch, after a feature drop, or when growth has stalled. The skill prompts for product context, audience, and current stage when missing, then narrows the catalog to a handful of high-fit tactics instead of dumping the full list. Categories include content and SEO plays such as easy keyword ranking, public SEO audits, and glossary marketing, plus broader growth mechanics documented in the full skill body. It is editorial and procedural knowledge packaged for agents, meant to turn vague “how do I market this?” requests into actionable next steps sized to indie constraints.656installs59Security ComplianceSecurity Compliance in this package is effectively an incident response report template for solo and indie teams who need a repeatable document when something goes wrong in production. Instead of improvising Slack threads into an audit trail, you get sections for incident ID, severity, status, commander, open/close timestamps, and a short executive summary. The body walks through classification (malware, phishing, breach, DDoS, unauthorized access), attack vector, affected assets, and a row-based timeline so every handoff—from SIEM alert through isolation, patching, forensics, and restore—is owned and dated. It fits builders shipping SaaS or APIs who may not have a full SOC but still need compliance-friendly records after ship. Use it during active incidents and for closed-loop postmortems; pair with your monitoring and security review skills rather than treating this as preventive scanning.656installs603d Web Experience3d-web-experience is an agent skill for solo and indie builders who want depth on the web without turning every page into a GPU trap. It positions you as a 3D web architect: pick the right stack, integrate models, and ship interactive scenes that still feel fast and usable. Use it when you are adding a Three.js or React Three Fiber scene, embedding Spline for a quick hero, or building a 3D product configurator on a SaaS or marketing site. The skill contrasts Spline, React Three Fiber, vanilla Three.js, and Babylon.js with a practical decision tree so you do not over-engineer a simple visual or under-power a complex one. It emphasizes visual impact balanced with accessibility for users who never touched a 3D app, plus WebGL optimization patterns so immersive moments do not tank mobile performance or Core Web Vitals.655installs61Excalidrawexcalidraw teaches coding agents a context-budget discipline: treat Excalidraw JSON as toxic to the main thread and route all opens, edits, and summaries through subagents. Solo builders sketching system design, onboarding flows, or integration maps in Excalidraw get accurate diagram work without losing half the window to strokeColor and seed fields. The skill is methodology-first—it does not replace Excalidraw the product but encodes when to spawn isolated workers for explain, update, and create requests. Journey-wide because the same trigger fires during Validate prototyping discussions, Build architecture docs, Ship runbooks, and Grow funnel diagrams. Pair it with repos that commit *.excalidraw alongside markdown. Beginner-friendly rule (“never read directly”) with intermediate payoff when you orchestrate multi-diagram PRs.650installs62PlotlyPlotly is a documentation-style agent skill that catalogs Plotly chart types and copy-pasteable Plotly Express patterns for solo builders who need readable analytics fast. It walks through basic charts—scatter, line, bar, pie, donut, and area—and statistical plots such as histograms with marginal box or violin views, box plots, and density heatmaps. Each section uses minimal pandas-friendly examples with color, size, orientation, and barmode parameters called out explicitly. Indie developers use it during Build when embedding charts in SaaS admin pages, internal tools, or Jupyter-style reports, and again in Grow when refining lifecycle or funnel visuals. The skill does not replace a full design system; it standardizes how an agent picks the right visualization primitive and Express API so you spend less time guessing px versus graph_objects syntax.640installs63Context7 Auto Researchcontext7-auto-research keeps your coding agent aligned with current framework documentation instead of stale training cutoffs. Solo builders installing skills from claude-code-templates use it whenever a task touches a fast-moving library—routing, data layers, auth SDKs, or UI kits—where guessing method signatures wastes ship time. The skill wraps the Context7 API with an auto-trigger path so documentation retrieval feels native to conversation flow rather than a manual browser tab ritual. Optional API key configuration improves throughput for heavier repos. It is procedural knowledge, not a hosted doc site: your agent still applies judgment, but arguments and imports are far less likely to be wrong-version. Reach for it across Idea research spikes, Validate prototypes, Build features, Ship fixes, and Operate patches whenever external package truth matters.639installs64Devops Iac EngineerDevOps IaC Engineer is an agent skill packaged as a DevOps teammate for indie builders shipping SaaS, APIs, or containerized services. It consolidates guidance for designing cloud architectures with Infrastructure as Code, deploying workloads on Kubernetes, choosing patterns across AWS, Azure, and GCP, and wiring CI/CD or GitOps flows that keep releases repeatable. The skill points to focused reference docs for Terraform practices, cluster operations, pipeline design, monitoring and logging, and DevSecOps controls so you do not stitch advice from scattered chats. Reach for it when you are standing up environments, hardening deploy paths, or improving observability after launch—not when you are only sketching a landing page. Complexity is intermediate: you should be comfortable with repos, containers, and basic cloud accounts. Outputs are architecture recommendations, IaC-oriented patterns, and pointers to template material rather than a fully provisioned account without your review.638installs65Statistical AnalysisStatistical Analysis is a reference-style agent skill that walks solo builders through validating statistical assumptions before interpreting test results. It encodes general principles—always check assumptions, combine visual and formal diagnostics, note robustness, and document remedial actions—and then details common failures such as non-independent observations and non-normal residuals. The material is aimed at indie founders and small teams who run their own A/B reads, funnel analyses, or research summaries without a dedicated statistician. Use it when you are about to claim significance, compare groups, or fit a linear model and need a disciplined checklist instead of trusting default library output. On Prism it sits in Validate because sound scope decisions depend on whether your design and data actually support the test you chose; it also supports Build and Grow when analytics code and reporting pipelines need the same rigor.637installs66Pubmed Databasepubmed-database packages Claude Code–friendly reference material for NCBI E-utilities so solo builders can pull PubMed literature without tab-hopping. It explains how to register an API key, attach it with &api_key, respect stricter rate limits when unauthenticated, and call esearch.fcgi (and related Entrez tools) with the right db, term, paging, and history-server options for big PMID sets. The skill is ideal when you are validating a health-tech idea, writing evidence-backed content, or building an agent tool that needs structured abstract metadata. It is documentation-first: you still implement fetch logic in your stack, but you avoid classic mistakes like missing User-Agent headers or blasting the API at three requests per second. Beginner-friendly for anyone who has made HTTP calls before; advanced biocuration pipelines may need fuller efetch/esummary orchestration beyond the excerpt.630installs67CopywritingCopywriting is an agent skill that turns you into a conversion copywriter for marketing pages. Solo and indie builders use it when they need homepage, landing, pricing, feature, about, or product page copy written from scratch or improved without guessing at positioning. The workflow starts by gathering page purpose, ideal customer, objections, offer differentiation, proof, and where traffic is coming from, then applies clear copy principles so headlines and CTAs drive one primary action. It fits whenever you are shaping how strangers understand your product, whether you are drafting a validate landing page, polishing launch campaign destinations, or refreshing grow lifecycle pages. Install it when chat drafts feel vague or off-brand; skip it for dedicated email sequences or popup experiments, which have sibling skills in the same template pack.629installs68Frontend Dev GuidelinesFrontend Dev Guidelines is a template-oriented agent skill from Claude Code Templates that encodes frequent UI patterns for React applications using MUI, React Hook Form, Zod, and a centralized useAuth hook. Solo builders use it while implementing dashboards, settings screens, and data entry flows so coding agents follow house style instead of sprinkling raw fetch calls for session state. The readme stresses never making direct API calls for authentication when useAuth is available, and shows complete form setups with schema validation and snackbar feedback. It targets TypeScript React codebases with path aliases like @/hooks and Material UI components. Complexity is intermediate because you need familiarity with hooks and resolver-based validation. It does not replace system design or API contract work on the backend; it keeps frontend agent output consistent during the build phase.623installs69Agent Memory SystemsAgent-memory-systems trains you and your coding agent to think like a cognitive architect: memory is what makes agents feel intelligent across turns and sessions. The skill spans short-term context limits, long-term vector storage, and how episodic, semantic, and procedural memories should be separated so the right fact surfaces at the right time. Its central thesis is that most "dumb" or inconsistent agent behavior is a retrieval and chunking problem, not missing disk space. You get patterns for picking vector databases, structuring memory types, and designing formation and decay so scale does not drown signal. Solo builders shipping Claude Code or Cursor agents with custom knowledge bases, support bots, or multi-step workflows should invoke this while designing Build agent-tooling, and revisit during Ship review or Grow lifecycle when user history must compound without blowing context windows.605installs70Pdf Processing ProPDF Processing Pro is a Claude Code Templates skill packaging production-oriented PDF tooling for solo builders who outgrew copy-paste pdfplumber snippets. It targets complex workflows—interactive forms, tabular reports, OCR-heavy scans, validation gates, and batch runs—where silent failures would cost support time or compliance headaches. Included scripts (analyze_form, fill_form, extract_tables) expose CLI entry points with logging, type hints, exit codes, and input validation so you can wire them into backends, cron jobs, or agent-driven refactor tasks. Use it when you are building document ingestion for a SaaS, automating contracts or invoices, or standing up ETL from PDF archives. It is not a hosted PDF API; you run Python locally or in your worker environment. After integration, plan Ship-phase testing on edge-case PDFs (encrypted, malformed, huge page counts) before exposing user uploads at scale.604installs71Claude Code GuideClaude Code Guide is a meta reference skill for solo builders who want consistent results from Claude Code across projects. It shows how to seed a root CLAUDE.md with run, test, and build commands plus style conventions, and how to steer deeper reasoning with explicit thinking keywords. The guide covers practical debugging when context goes stale or instructions are misapplied, emphasizing read-first discipline and post-edit verification. Because it addresses configuration and interaction patterns rather than a single feature, it remains useful whenever you onboard a new repo, tune agent behavior, or recover from a confused session—without replacing domain-specific implementation skills.600installs72Perplexity Searchperplexity-search is a configuration and model-selection guide for using Perplexity’s Sonar family via OpenRouter inside Claude Code template workflows. Solo builders set OPENROUTER_API_KEY and optional defaults, then pick a model tier from quick lookups to agentic multi-step research. The readme contrasts Sonar Pro Search for comprehensive synthesis against lighter models for faster answers. It suits early opportunity research, competitive landscape scans, and technical deep dives where you want cited, search-augmented responses instead of static training data alone.600installs73Clinical ReportsClinical Reports from davila7’s claude-code-templates is a documentation skill that walks an agent through a publishable clinical case report. Solo clinicians, residents, and indie health writers use it when they have an interesting patient story and need IMRAD-adjacent structure without guessing journal expectations. The template enforces familiar front matter—title conventions, author block, MeSH-style keywords—and a structured abstract with explicit subsection prompts and a 150–250 word target. The introduction section is decomposed into background, novelty, and objectives paragraphs so the agent does not skip why the case matters. Patient information and subsequent clinical sections (truncated in the catalog excerpt) continue the same pattern. Pair it with your institution’s ethics and de-identification rules; the skill organizes prose but does not replace IRB, consent, or PHI policies.592installs74Ai Agents ArchitectAI Agents Architect is an agent skill that positions your coding assistant as a systems architect for autonomous agents. It walks solo and indie builders through architecture choices that stay controllable when models misfire: ReAct reason-act-observe loops with max iterations, plan-then-execute with optional replanning, and centralized tool registries. It also covers memory design, multi-agent orchestration, and how to evaluate and debug agent runs. Use it when you are starting a new agent feature, refactoring ad-hoc tool scripts into a coherent loop, or hardening production agents that currently fail in opaque ways. The skill is procedural knowledge for Claude Code, Cursor, and Codex—not a hosted runtime. Expect intermediate depth: you should already use an LLM API and understand basic function calling and prompting.591installs75Content CreatorContent Creator is a structured content-calendar template your agent can instantiate for a given month and year. It is aimed at solo and indie builders who publish across several channels but do not have a dedicated marketing team. The layout forces explicit monthly goals, then walks week by week through concrete placements—long-form blog posts, LinkedIn shares, Instagram carousels, segmented email sends, Twitter/X threads, and coordinated Friday campaigns that list every asset you still owe. Each slot includes keywords, captions, hashtags, timing, and simple status toggles so an agent can draft copy, mark progress, or surface gaps before publish day. A rolling content bank captures ideas when you are not ready to schedule them, and the performance-review section closes the loop with top posts, qualitative wins, and adjustment notes for the next cycle. Use it whenever you need a repeatable editorial spine instead of improvising posts in chat.588installs76Browser AutomationBrowser Automation is an agent skill for solo builders who need dependable control of real browsers—whether for launch-blocking e2e suites, maintainable scrapers, or agents that click through authenticated web apps. It treats Playwright as the default framework in modern stacks, reserving Puppeteer for niche stealth or Chrome-only constraints, and trains you to avoid the three classic failure modes: brittle selectors, missing waits, and bot detection on scraping paths. Patterns emphasize isolated tests, user-visible locators, and auto-wait instead of timing guesses, which shrinks flake rates as products grow. Use it while hardening CI before Ship, when adding regression coverage to a solo SaaS, or when building agent-tooling that must navigate dashboards on behalf of users. The skill does not replace a full observability stack; it gives procedural knowledge your coding agent can apply to scripts, fixtures, and CI jobs. Expect to still configure secrets, base URLs, and storage state yourself—the skill encodes how experts structure automation, not your vendor-specific hosting.587installs77Api Integration SpecialistAPI Integration Specialist is an agent skill for solo builders who need external APIs wired correctly the first time—authentication, error handling, backoff, and data mapping included. It targets REST and GraphQL consumers, webhook receivers, and thin client wrappers around services like billing, messaging, or email providers. The skill encodes security habits such as environment-stored keys, structured OAuth client configuration, and scoped credentials rather than hard-coded secrets. Operational concerns—rate limits, retries, circuit breaking, and debugging failing calls—are treated as part of integration design, not afterthoughts. Use it when your agent is generating or fixing vendor SDK and HTTP client code. It spans Build integrations as the primary shelf and Ship or Operate when you are stabilizing production traffic or diagnosing integration failures.584installs78Frontend PatternsFrontend Patterns is a reference-style agent skill for solo and indie builders shipping React or Next.js products. It encodes modern UI architecture—component composition, compound components, state management, and performance habits—as copy-paste TypeScript examples your agent can follow instead of inventing one-off structures. Install it when you are building dashboards, marketing sites, or SaaS shells and want consistent Card/Tab patterns, context-based APIs, and performance-minded defaults across sessions. It does not replace a design system or Figma handoff; it standardizes how your agent writes frontend code. Because patterns apply whenever you edit UI—from a validate prototype to a grow-phase settings page—journey scope is journey-wide with a Build → frontend canonical shelf for catalog browsing.582installs79Pytorch LightningPyTorch Lightning Callbacks is an agent skill focused on Lightning’s callback system so solo builders can extend training runs without entangling checkpointing and early stopping inside the model class. The documented flow centers on ModelCheckpoint: choosing metrics to monitor, min versus max modes, retaining the top K checkpoints, naming files with epoch and metric tokens, and loading the best weights back into a LitModule. It targets indie ML developers who need reliable experiment artifacts before they wire inference APIs or batch jobs. Callbacks capture ancillary training behavior—persistence, scheduling saves on validation end, verbosity—while the LightningModule stays focused on forward and loss logic. Invoke the skill when you are standing up a Trainer for the first time or when save/resume behavior is wrong and you want the agent to propose Lightning-native patterns instead of hand-rolled torch.save loops.578installs80Nestjs ExpertNest.js Expert is a framework-focused agent skill for solo builders shipping TypeScript APIs on Nest. It expects the agent to inspect the repo first, map existing modules, and apply Nest conventions for dependency injection, cross-cutting middleware, guards, interceptors, and validation pipes. The description emphasizes proactive use for architecture choices, performance tuning, and tricky provider graphs, plus structured testing with Jest and Supertest after typecheck. Integrations cover TypeORM and Mongoose as well as Passport-based auth. When a problem is really pure types, SQL tuning, raw Node runtime, or React UI, the skill instructs switching subagents instead of bluffing. That makes it a durable backend copilot across initial implementation and later hardening passes.576installs81Mermaid Diagramsmermaid-diagrams is a comprehensive agent skill for creating professional software diagrams from Mermaid’s text syntax. Solo builders use it whenever complex technical ideas need to become clear, maintainable visuals that stay in git beside the code. The skill walks through picking the right diagram type—architecture context, containers, class and domain models, sequence and API interactions, entity-relationship database schemas, flowcharts, and user journeys—and producing definitions teams can update in documentation and code review. It fits early alignment during validation, ongoing build-time documentation, and ship-phase explanations for reviewers. Because output is plain text, diagrams evolve with the product instead of rotting in static image files. Intermediate familiarity with Mermaid block types and naming conventions helps; the skill is less about pixel polish and more about accurate structure and communication.570installs82Agent EvaluationAgent Evaluation packages quality-engineering judgment for LLM agents: the same prompt can yield different valid answers, so pass/fail cannot mirror traditional unit tests naively. The skill steers solo builders toward statistical reruns, behavioral contracts, capability assessments, and adversarial probes rather than chasing a mythical 100% green dashboard. It references real-world benchmark gaps—where top agents still fail on production-like tasks—to set expectations before you bet a product on brittle automations. Invoke it when you are adding skills, MCP tools, or autonomous loops and need regression coverage that survives model updates. It assumes familiarity with testing fundamentals and LLM behavior, and it complements manual code review by focusing on agent trajectories, tool use, and reliability metrics. Use outcomes to decide ship/no-ship and to define monitors once agents run against live users.564installs83Nextjs Best Practicesnextjs-best-practices is an agent skill that encodes Next.js App Router principles for solo builders shipping React apps with Claude Code, Cursor, or similar agents. It walks through when to default to Server Components versus marking interactive UI with use client, how to choose caching and revalidation for database and API data, and how to organize routes with layouts and loading states without fighting the framework. The guidance assumes you are actively building pages and features—not doing market research or ops monitoring—and is meant to reduce rework from mixing client-only patterns into server trees or fetching dynamically when static would suffice. Install it when your agent keeps generating Pages Router habits or ambiguous fetch/cache directives on App Router projects.562installs84Peer ReviewPeer review is an agent skill that equips solo researchers, indie science writers, and builder-founders publishing technical reports with a structured catalog of methodological and statistical issues seen in real manuscript reviews. Use it when a draft claims significance, equivalence, or causal effects and you need an agent to behave like a careful reviewer rather than a cheerleader. The material spans p-hacking signals, uncorrected multiple comparisons, parametric tests on violated assumptions, and the habitual conflation of statistical and scientific significance. Because those failures show up while you are still shaping a study narrative and again right before you ship a PDF or preprint, the skill is multi-phase even though Prism shelves it under Ship review. It produces issue-focused, actionable commentary you can fold into revision notes or response-to-reviewers outlines. It does not run analyses or fetch literature; it pattern-matches prose and methods sections against known failure modes and suggests fixes aligned with open-science norms.553installs85Prompt Engineeringprompt-engineering is an expert agent skill for advanced LLM prompt design: few-shot exemplars, chain-of-thought, and optimization patterns that improve reliability and controllability. Solo and indie builders shipping with Claude Code, Cursor, or Codex install it when model outputs are inconsistent, JSON or schemas are wrong, or they want to learn systematic prompting instead of one-off chat tweaks. The skill teaches when to trade tokens for more examples, how to show reasoning before answers on analytical work, and how to mirror domain tasks (e.g. support-ticket extraction) in markdown prompt blocks. It is procedural knowledge—usable in Idea research summaries, Validate prototypes, Build features, Ship reviews, and Operate incident triage whenever an agent is in the loop. It complements skill packs and MCP tools by improving the instructions layer, not replacing integrations.543installs86Cto Advisorcto-advisor packages an Architecture Decision Records (ADR) framework for solo builders and tiny teams who need institutional memory without a formal architecture review board. When you are choosing monolith vs services, databases, or deployment models, the skill walks you through context, decision drivers, considered options, chosen outcome, and documented tradeoffs in a repeatable markdown template. It fits the Validate phase when scoping a product, the Build phase when backend structure solidifies, and Operate when you revisit or deprecate earlier choices. Indie developers using Claude Code benefit because agents can draft ADRs that stay consistent with standard sections (deciders, consequences, links to PoCs). The output is not a mandate to microservices—it is a decision log that makes onboarding, audits, and refactors faster because the why is searchable next to the code.542installs87Latex PostersLatex-posters is an agent skill that ships a structured baposter LaTeX template for research-style posters with automatic column layout and professional defaults. Indie builders and researchers who need a credible physical or PDF poster for a meetup, university demo, or conference booth can point their coding agent at this skill to fill in sections rather than researching poster class options from zero. The template targets a0 portrait, three-column flows, and common math and figure packages so agents focus on your narrative, results, and QR links. It fits the Validate-to-Build stretch when you are turning a prototype or study into something you can hang on a wall or share as a large-format PDF. Expect intermediate LaTeX comfort or an agent that can compile and iterate on missing packages locally.542installs88Web Design GuidelinesWeb-design-guidelines is an agent skill that turns Vercel’s Web Interface Guidelines into a repeatable UI compliance review for solo builders shipping web products with Claude Code, Cursor, or Codex. When you ask to review my UI, check accessibility, audit design, review UX, or check a site against best practices, the skill first retrieves the current rule set from the official raw command.md source, then reads the paths or patterns you supply, and applies every rule from that snapshot. It is built for builders who already have markup or components and want an objective pass before merge or launch, not for open-ended visual exploration. Findings are reported in the compact file:line style the upstream guidelines specify, which makes fixes easy to assign in your editor or agent loop. Because rules are fetched fresh each time, audits stay aligned with guideline updates without maintaining a local copy in the repo.542installs89Product Manager ToolkitProduct Manager Toolkit is an agent skill packaged as procedural PRD templates for solo and indie builders who act as their own PM. It walks you through an executive summary stakeholders can scan in one page, then deepens into customer problem definition, market opportunity, and a business case so you are not coding on vibes alone. The solution section forces a clear user journey, differentiated value proposition, and explicit in-scope versus out-of-scope feature lists with priorities—reducing rework when you hand work to Claude Code, Cursor, or Codex. Use it in Validate when you are scoping an MVP, in Idea when you need to articulate market timing after research, or in Build when you are refreshing specs mid-sprint. The skill outputs structured markdown-ready sections rather than integrations or live analytics; you bring the customer insight and numbers. Complexity is beginner-friendly if you already know your user and problem hypothesis.540installs90Agent Manager SkillAgent Manager Skill is a procedural workspace package for solo and indie builders who run more than one local CLI coding agent at once. It wraps a Python CLI (`agent-manager/scripts/main.py`) around tmux so each agent gets an isolated session you can start, stop, list, tail with `--follow`, and feed assignments without juggling terminals manually. Typical flow: clone the repo, configure agents under `agents/`, run `doctor`, then `start EMP_0001` and `monitor` while you build features in parallel. The assign command pipes markdown workflows (for example maintenance playbooks) into a chosen agent. Prerequisites are tmux and Python 3; agents are file-based configs in the cloned skill tree. Use when parallel agent throughput matters more than a single chat thread—maintenance sweeps, multi-repo chores, or scheduled cron batches—not when you only need one interactive Claude Code session.536installs91Product StrategistProduct Strategist is an agent skill oriented around an OKR Cascade Generator: procedural help for turning high-level company strategy into aligned objectives and key results at team level. Solo founders and tiny teams can use it when strategy lives in their head but execution needs numbered outcomes—growth templates cover acquisition, market expansion, and CAC efficiency, while retention templates emphasize churn, NPS, and stickiness with placeholders for current and target values. The embedded Python-oriented reference models how objectives pair with reusable key-result phrasing so you are not staring at a blank OKR doc each quarter. It fits early validation when you are scoping what success means before sprinting, and again in build and grow phases when you rebalance metrics after shipping. Treat it as planning scaffolding the agent fills in with your real numbers, not a substitute for customer interviews or financial modeling.535installs92Agent Memory McpAgent Memory MCP is an integration skill that boots a searchable, persistent memory server for AI coding agents. Solo builders shipping agent-heavy products use it in Build to stop re-explaining architecture every session. Setup clones the agentMemory repo into the skills workspace, installs Node dependencies, compiles, and starts the MCP server scoped to a project id and absolute workspace path. Exposed tools let the agent query prior patterns (“authentication”) and record new decisions with structured types and tags. The model fits multi-week builds and later Operate work where prior decisions must stay discoverable. It requires local Node tooling and comfort running MCP alongside your editor. It complements ad-hoc markdown notes by making memory machine-retrievable through standard MCP calls rather than brittle copy-paste context.532installs93I18n LocalizationThe i18n-localization skill packages a small Python scanner that audits your frontend and backend source for strings that should live in translation catalogs instead of hardcoded literals. Solo builders shipping SaaS or APIs with React, Vue, or Python surfaces invoke it when adding locales or cleaning up tech debt before release. It contrasts suspicious capitalization-heavy literals against established i18n call patterns so your agent can prioritize refactors. It is a focused checker—not a full TMS integration—meant to complement react-i18next, vue-i18n, or gettext workflows. Run it after feature work or before merge when multilingual support is on the roadmap, and feed findings into your normal PR workflow.525installs94Feature Design AssistantFeature Design Assistant is a workflow skill that guides an agent through discovery on your real repo, then batches structured questions to collect goals, users, constraints, and scope before producing a design collaborators can approve. It targets solo builders planning net-new capabilities, major enhancements, or architectural shifts—not drive-by one-liners. The flow explicitly opens with a skill announcement, explores existing patterns to avoid duplicate systems, and uses multi-round questioning instead of guessing requirements. That reduces rework when Claude Code or Cursor would otherwise codegen against an ambiguous prompt. Place it early in Validate or at the front of Build PM work; hand off to implementation-plan or coding skills once the spec is stable.523installs95Seo Fundamentalsseo-fundamentals packages a repo-local SEO checker aimed at solo and indie builders who ship marketing sites, SaaS landings, or React app shells and need confidence that pages are crawlable and shareable before launch. The skill drives a Python script that walks your project tree, ignores build artifacts and non-page code, and flags gaps in meta tags, Open Graph data, semantic heading structure, and accessible image alt text on HTML and JSX/TSX files likely to be public routes. It complements agent-assisted development by turning vague “check SEO” requests into a repeatable audit you can run after refactors or before a release. Use it in the Launch phase when routes are mostly stable; pair results with content edits and sitemap or robots work elsewhere in your stack. It does not replace keyword strategy, Core Web Vitals tuning, or server-side rendering audits—it is a fast front-of-house markup and structure pass.522installs96SeabornSeaborn is a Python agent skill that teaches practical, copy-paste visualization workflows on top of pandas, matplotlib, and numpy. Solo builders use it when they need credible charts during exploration, feature engineering, or stakeholder-facing reports without relearning seaborn’s API surface each time. The readme walks through exploratory data analysis (pairplots with hue and corner layouts, faceted kde displots), correlation matrices with upper-triangle masks, and scientific multi-panel figure setups. It emphasizes sensible defaults—hue by target or condition, col_wrap faceting, common_norm control, and export discipline—so agents produce consistent filenames and DPI rather than inline-only plots. Install it when your agent is writing analysis notebooks, validating datasets before modeling, or embedding visualization snippets into backend jobs that emit PNGs. It is procedural knowledge, not a running service: you still need your CSV or DataFrame in the repo and matplotlib configured in the environment.520installs97Tailwind PatternsTailwind-patterns is an agent skill that teaches solo builders how to work in Tailwind CSS v4 using CSS-native configuration, modern layout primitives, and opinionated component class recipes. It explains what changed from v3—@theme instead of JavaScript config, the Oxide compiler, exposed design tokens as CSS variables—and walks through when to extend versus override theme values, how to wire semantic color and spacing scales, and how to structure cards, buttons, and forms with consistent states. Container queries, a mobile-first responsive ladder, class-based dark mode, and performance habits (content paths, avoiding @apply churn) are spelled out so agents do not recreate desktop-first or magic-number layouts. Install it when you or your coding agent are building or refactoring a web UI on Tailwind v4 and need repeatable patterns rather than one-off class soup.519installs98ArchitectureArchitecture is a journey-wide agent skill from Claude Code templates that stops your coding assistant from jumping to microservices or framework buzzwords without context. It instructs the agent to interview you first on scale, team size, timeline, domain complexity, and constraints, then maps answers onto an MVP versus SaaS versus Enterprise matrix with example stacks such as simple Next.js APIs through NestJS and distributed services. Solo and indie builders install it to turn vague “how should I build this?” chats into scoped recommendations aligned with actual user counts, data volume, and time pressure. The skill is procedural knowledge—not a deploy integration—so it fits Validate scoping, Build planning, and Ship conversations whenever you refactor or scale. It matters because wrong early architecture burns months; the skill encodes a repeatable discovery ritual and concrete classification examples so agents propose patterns you can defend in a one-person roadmap.515installs99Brainstormingbrainstorming is a journey-wide agent skill from the davila7 Claude Code templates that forces a design conversation before your agent writes code. It treats implementation as the last step: first align on project context, then refine the idea through focused questions—one per turn—about purpose, constraints, and success criteria. When the problem space is clear, it explores two or three architectural approaches with explicit trade-offs and a stated recommendation, then drafts the design in digestible sections so you can correct course early. Solo builders benefit because it prevents expensive rework on features, UI components, integrations, and behavioral tweaks across any stack. Use it whenever the spec lives only in chat, when multiple valid designs exist, or when you want the agent to behave like a patient product partner instead of jumping straight to patches.514installs100Research Lookupresearch-lookup is an agent-oriented wrapper around a ResearchLookup client that routes natural-language research questions to the right Sonar-tier model instead of forcing one model for every prompt. Solo builders use it when validating hypotheses, drafting technical content, or preparing specs that need cited, up-to-date findings without hand-picking APIs on each turn. The bundled examples show automatic selection for straightforward lookups versus comparative analysis, plus optional overrides when you want cheaper fast search. It fits early discovery when you are still deciding what is true in the market or literature, and it remains useful during Build and Launch when documentation, changelogs, or positioning need refreshed citations. Install it when your coding agent should behave like a disciplined research assistant rather than improvising facts from training data alone.513installs101PineconePinecone is an agent skill that teaches solo builders how to deploy and operate Pinecone vector indexes in production using the official Python client. It walks through creating serverless indexes with ServerlessSpec across AWS, GCP, or Azure regions—ideal when traffic is spiky and you want auto-scaling without managing pods. It also covers pod-based PodSpec setups with pod types, replica counts, and dedicated capacity when you need stable latency and higher query throughput. The guide adds hybrid search by upserting dense embeddings alongside sparse token indices so retrieval can blend semantic similarity with lexical signals. Use it while wiring RAG, agent memory, or product search in build and integrations, then revisit at ship and operate when you promote an index from dev keys to real traffic, regions, and cost controls.512installs102PolarsPolars is an agent skill that teaches solo builders how to write efficient Polars code for real datasets: prefer lazy evaluation, push filters and projections early, and avoid Python functions that break vectorization and parallel plans. It is for anyone shipping Python backends, CLI data tools, or agent workflows that read CSV/Parquet, join tables, and aggregate metrics without drowning in memory or wall time. Use it when you or your coding agent is drafting transforms, reviewing a slow pipeline, or refactoring pandas-style eager scripts into production-grade Polars. The guide walks through concrete bad-vs-good patterns—eager reads versus scan plus collect, late filters versus early select/filter, and expression-native logic versus per-row Python callbacks—so outcomes are faster IO, smaller intermediate frames, and query plans the engine can optimize. It fits the Build phase as procedural knowledge you invoke during implementation and refactors, not as a launch or growth play.512installs103Audiocraft Audio GenerationAudioCraft Audio Generation is an agent skill for builders who want controllable music from Meta’s AudioCraft stack, especially MusicGen fine-tuning on their own clips. It walks through preparing a labeled audio directory: load files, resample to 32 kHz, normalize to mono, emit consistent filenames, and build metadata JSON descriptions that trainers expect. Solo and indie developers use it when a generic model does not match a brand sound, game loop, or product demo and they need a reproducible dataset pipeline instead of ad-hoc exports. The skill fits the Build phase as backend ML prep you run before integrating inference into an API, agent tool, or content workflow. Expect intermediate comfort with Python, PyTorch/torchaudio, and GPU time. It does not replace hosted music APIs; it documents how to structure data and processing for serious fine-tuning work inside a repo.510installs104Seo AuditSEO Audit is an agent skill that walks you through a search-focused diagnostic when you ask for an SEO audit, technical SEO help, or wonder why you are not ranking. It starts with context—site type, goals, priority keywords, known migrations, and whether you can share Search Console—then works in a fixed priority order: crawlability and indexation, technical foundations, on-page optimization, content quality, and authority. Solo builders use it before a launch push, after a redesign, or when organic traffic stalls, because it turns vague “fix my SEO” requests into actionable recommendations instead of generic keyword tips. The skill points to sibling flows for programmatic page building and schema markup when those are the real bottleneck. It is methodology and review oriented, not a connector that changes your CMS for you.509installs105Marketing Demand AcquisitionMarketing Demand Acquisition is a lightweight agent skill centered on a Python CAC (Customer Acquisition Cost) calculator for solo founders validating whether marketing spend is efficient. Install it when you have campaign spend and customer counts by channel—or only totals—and need blended and channel-specific CAC without spinning up a full analytics stack. The workflow accepts structured channel dictionaries, computes per-channel and blended figures, and formats output for growth reviews, pricing sanity checks, or board-style summaries. It suits indie SaaS, content products, and small e-commerce shops comparing ads, SEO-assisted signups, or partnerships in one view. It does not replace attribution tooling or cohort LTV analysis; it gives fast, explicit math your agent can re-run as you update monthly numbers.504installs106Agent Tool Builderagent-tool-builder is an agent skill for solo builders shipping AI products who need tools that work on the first call—not loops, silent failures, or mystery 10x token bills. It treats the schema and natural-language description as the real API surface, because models never execute your Python; they only choose arguments against what you documented. You get patterns for crisp JSON Schema, examples that steer usage, validation boundaries, and errors that teach the model how to retry. MCP is framed as the practical lingua franca when you want the same tool usable across Claude Code, Cursor, and other hosts. Use it while defining new capabilities in Build, and again in Ship or Operate when tightening production tool contracts after real agent traces.502installs107Notion Knowledge CaptureNotion Knowledge Capture is an agent skill for solo builders who lose decisions in chat threads and need a repeatable path into Notion. It walks you from clarifying audience and capture type through picking the right database template, pulling existing context with Notion search and fetch, drafting structured pages with summaries, source links, tags, and owners, then linking hubs and updating status as sources evolve. The workflow assumes Notion MCP connectivity and documents a pause-and-setup sequence—adding the remote MCP, enabling rmcp_client, and completing OAuth—when calls fail. Multi-phase by nature: you capture during Build docs sprints, Grow content operations, and Operate iteration when incidents spawn new runbooks. Best for agent-forward teams already on Notion; skip if you refuse cloud knowledge bases or cannot enable MCP OAuth in your Codex environment.502installs108Gdpr Dsgvo Expertgdpr-dsgvo-expert is a senior-level agent skill for solo and indie builders who process personal data in the EU or Germany and need structured GDPR and DSGVO guidance instead of vague “be compliant” chat advice. It is invoked for compliance assessments, privacy audits, data-protection planning, and verification against Articles covering lawful basis, special categories, consent, information duties, access, rectification, erasure, restriction, portability, and objection. The skill encodes framework-style checklists and auditor thinking so your agent can draft gap analyses, DPIA outlines, and governance recommendations aligned with accountability obligations. It does not replace licensed legal counsel for binding interpretations, but it compresses expert audit methodology into repeatable agent sessions during Validate scoping and Ship security hardening.499installs109React Best PracticesReact Best Practices is an agent skill that packages a Vercel Engineering performance guide for React and Next.js into actionable rules a solo builder can invoke while optimizing or reviewing an app. It targets developers who already ship with React 19-era patterns and need a structured checklist instead of scattered blog posts. The skill is meant when you are debugging slow loads, shrinking bundles, refactoring components, or implementing new UI with performance guardrails from day one. Rules span eliminating request waterfalls, cutting bundle size, tuning React Server Components and data fetching, improving client-side caching, and reducing unnecessary re-renders. For Prism’s journey model it shelves under Ship → Performance as the canonical home, with strong overlap in Build → Frontend whenever you write or refactor components. It pairs well with code review and testing skills but does not replace profiling tools or automated Lighthouse runs—you still validate in the browser and CI.477installs110Api Documentation Generatorapi-documentation-generator is a Build-phase generator skill for solo builders who ship APIs faster than they write manuals. It walks the agent through analyzing your codebase—routes, HTTP verbs, request and response shapes, auth, and errors—then emits structured endpoint documentation with copy-paste examples and usage guidelines. It is aimed at REST, GraphQL, and WebSocket surfaces when docs are missing, outdated, or blocking onboarding. You invoke it when a feature set is stable enough to describe but you do not want to hand-maintain every parameter table. Output is meant to be developer-friendly and suitable for tightening into OpenAPI or Swagger specifications. It complements code review and security passes but does not replace contract tests or live gateway configuration.474installs111Isms Audit ExpertISMS Audit Expert is listed in davila7’s Claude Code templates as a security-focused agent skill aimed at Information Security Management System audits—the kind of structured control review aligned with ISO 27001 thinking. Prism’s ingest shows placeholder asset and reference readme sections rather than a complete procedural SKILL.md, so treat this entry as a catalog stub you may need to flesh out locally before production audits. For solo founders selling B2B SaaS, the intended use is to walk gaps in policies, risk treatment, access control evidence, and logging retention with consistent questioning, then map answers to control objectives. It is phase-specific to Ship/security because audits gate launch and enterprise deals more than day-one coding. Confidence is moderated until real checklist steps and output formats appear in the repository.1installs