Skillstack
Browse and install a large curated skillstack of engineering and meta-infra Claude plugins—evaluation, API design, agent project pipelines, and more.
Overview
skillstack is a plugin marketplace for the Build phase that distributes 66 engineering and meta-infra Claude skills including agent evaluation, API design, and LLM project development.
What is this marketplace?
- 66 plugins in one marketplace from Viktor Bezdek’s skillstack catalog
- agent-evaluation: rubrics, LLM-as-judge, pairwise comparison, bias mitigation, monitoring
- agent-project-development: task-model fit, acquire-prepare-process-parse-render pipelines, cost estimation
- api-design: REST, GraphQL, gRPC, and OpenAPI-oriented comprehensive API design skill
- Categories span Meta-Infra and Engineering for agent systems and product APIs
- 66 plugins listed in the skillstack marketplace catalog
- Representative plugins include agent-evaluation v1.0.5, agent-project-development v1.0.5, and api-design v1.2.24
Community signal: 9 GitHub stars.
What problem does it solve?
You need serious agent-engineering and API skills in Claude Code but do not want to hunt GitHub one repo at a time or maintain forks yourself.
Who is it for?
Indie builders shipping agent-powered products who want a broad, versioned skill library under one marketplace namespace.
Skip if: Beginners who want one minimal lifecycle plugin only, or teams that do not use Claude Code’s plugin ecosystem.
What do I get? / Deliverables
You install targeted skillstack plugins—evaluation, API design, project pipelines—so Claude sessions follow structured rubrics and methodologies out of the box.
- Installed domain skills (e.g., API specs, eval rubrics, project pipelines) in your agent
- Versioned plugin sources you can update per skillstack releases
- Repeatable Claude workflows aligned to each chosen plugin’s methodology
Plugins in this marketplace
66 plugins — install individually after you add the marketplace.
PluginVersion
Agent EvaluationComprehensive evaluation framework for LLM agent systems. Multi-dimensional rubrics, LLM-as-judge with bias mitigation, pairwise comparison, direct scoring, confidence calibration, and continuous monitoring.1.0.5Agent Project DevelopmentMethodology for LLM-powered project development. Task-model fit analysis, pipeline architecture (acquire-prepare-process-parse-render), file system state machines, cost estimation, and architectural reduction.1.0.5Api DesignComprehensive API design skill for REST, GraphQL, gRPC, and Python library architectures. Design endpoints, schemas, authentication, pagination, error handling, and federation.1.2.24Bdi Mental StatesBelief-Desire-Intention cognitive architecture for LLM agents. Formal BDI ontology, T2B2T paradigm, RDF integration, SPARQL competency queries, and neuro-symbolic AI integration patterns.1.0.5Brainstorm SwarmRun a parallel persona-swarm interview to brainstorm any complex decision — feature design, architecture choice, content piece, organizational change, product strategy. Spawns 6-12 persona-distinct subagents in parallel (PM, Engineer, Designer, Skeptic, User Advocate, Pre-Mortem Specialist, Junior, Veteran, First-Principles Thinker, Constraint-Setter, Optimist, Operator), each interviewing the user from their perspective and contributing questions, concerns, and ideas. Synthesizes the multi-perspective output into consensus, dissent, and open-questions. Use when the user asks to brainstorm with multiple perspectives, run a persona swarm, get a virtual roundtable, workshop an idea from PM/engineer/designer/skeptic angles, pre-mortem a decision, or think through something from different angles. NOT for code review (use code-review). NOT for single-perspective interviews (use elicitation or deep-interview). NOT for executing or building (use team, autopilot, or multi-agent-patterns). NOT for creating product personas as artifacts (use persona-definition). NOT for stakeholder mapping (use persona-mapping).1.1.0Cicd PipelinesComprehensive CI/CD pipeline design, DevOps automation, infrastructure as code, GitOps deployment automation, security scanning, and enterprise pipeline readiness.1.1.24Cloud InfrastructureInfrastructure-as-code with Terraform and AWS CDK, cloud architecture patterns, cost optimisation, and multi-region deployment design.1.0.0Cloud FinopsExpert Cloud FinOps guidance covering AI cost management, GenAI capacity planning, AI-powered FinOps automation, cloud billing (AWS, Azure, GCP, OCI), Kubernetes/container FinOps, serverless FinOps, data platform FinOps (Kafka, OpenSearch, Redis/Valkey), multi-cloud normalization (FOCUS spec), commitment strategy, tagging governance, SaaS asset management, ITAM, and GreenOps. Includes 26 domain-specific reference files grounded in enterprise delivery experience. Built by OptimNow and Viktor Bezdek, licensed CC BY-SA 4.0.2.1.2Code ReviewPerform thorough code reviews with multi-agent swarm analysis covering security, performance, style, tests, and documentation. Analyze PRs, extract and prioritize comments, and generate actionable fix plans.1.1.25Coding DisciplineResearch-grounded 5-principle behavioral contract for production LLM coding agents. Addresses named failure modes (test-gaming, phantom changes, scope creep, ego-signaling, confident hallucination) with empirical backing from SWE-bench, FeatBench, DELEGATE-52, and SE literature. Covers Think Before Coding, Simplicity First, Surgical Changes, Goal-Driven Execution, and Calibrated Communication — each with named anti-patterns, iteration budgets, and a structured completion schema.1.0.0CommunicationFive composable communication skills for people who write for work: structure a written piece (BLUF, Minto Pyramid), edit for clarity and conciseness (active voice, hedge removal), align stakeholders (RFCs, proposals, decision docs with DACI/RAPID roles), apply documentation discipline (ADRs, runbooks, decision logs), and communicate visually (Mermaid flowcharts, sequence diagrams, C4, diagram-as-code).1.0.1Competitive IntelligenceCompetitor analysis, market sizing, positioning maps, win/loss pattern synthesis, and battlecard creation for product and GTM decisions.1.0.0Consistency StandardsEstablish and maintain naming conventions, taxonomy standards, style guides, and reuse patterns across documentation and code.1.0.11Content ModellingDesign content models with types, fields, relationships, and governance rules for structured content systems.1.0.12Context CompressionProduction strategies for compressing LLM context windows. Anchored iterative summarization, opaque compression, tokens-per-task optimization, and probe-based evaluation.1.0.11Context DegradationPatterns for recognizing and mitigating context failures in LLM agents. Covers lost-in-middle, context poisoning, distraction, confusion, clash, and model-agnostic degradation measurement.1.0.11Context FundamentalsFoundational understanding of context engineering for AI agent systems. Covers context anatomy, attention mechanics, progressive disclosure, context budgeting, and the quality-vs-quantity principle.1.0.6Context OptimizationTechniques for extending effective context capacity through compaction, observation masking, KV-cache optimization, and context partitioning. Double or triple effective context without larger models.1.0.11Creative Problem SolvingGenerate breakthrough solutions through lateral thinking, first principles reasoning, game theory, and strategic reframing.1.0.16Critical IntuitionDetect hidden patterns, expose blind spots, and deliver rigorous critical analysis with intuition-level depth.1.0.16Database DesignSQL schema design, ORM patterns, migration strategies, query optimisation, and data-modelling for relational databases.1.0.0DebuggingComprehensive debugging skill combining systematic debugging methodology, browser DevTools automation, E2E testing with visual analysis, CI/CD pipeline debugging, performance profiling, and AI-powered error analysis.1.1.27DeslopRemove AI slop from UI copy, marketing text, and product content. Three focused skills: UI microcopy cleanup, marketing and editorial humanization, and AI-detection prevention.1.0.0Docker ContainerizationComprehensive Docker and containerization skill covering Docker basics, multi-stage builds, Docker Compose orchestration, container optimization, development environment setup, and infrastructure patterns.1.1.23Documentation GeneratorGenerate repository documentation at the right scale, from lightweight README updates to full codebase documentation sets with API, architecture, quality, and gaps coverage.1.2.1Edge Case CoverageIdentify and document boundary conditions, error scenarios, corner cases, and validation requirements.1.0.11Email MarketingEmail newsletters, drip sequences, subject line optimisation, onboarding emails, and plain-language transactional email copy.1.0.0ElicitationPsychological elicitation and deep-interview design using narrative identity (McAdams), self-defining memories (Singer), Motivational Interviewing (OARS), values elicitation (Schwartz), schema detection (Young), life review (Haight/Birren), and linguistic analysis (Pennebaker/LIWC). Progressive-disclosure structure with 8 domain references.2.0.1Example DesignDesign effective code examples, tutorials, and runnable samples with progressive complexity.1.0.11Filesystem ContextFilesystem-based context engineering patterns for LLM agents. Scratch pads, plan persistence, sub-agent communication, dynamic skill loading, terminal persistence, and self-modification patterns.1.0.14Frontend DesignComprehensive Frontend Design (UI/UX) skill combining UI design systems, component libraries, CSS/Tailwind styling, accessibility patterns, and visual design.1.1.24Git WorkflowComprehensive Git workflow management skill covering conventional commits, commit quality analysis, intelligent file grouping, worktree management with GitFlow conventions, issue tracking integration, changelog generation, semantic versioning, and hierarchical story backlog management.1.1.21Gws CliGoogle Workspace CLI (gws) skill for managing Drive, Gmail, Sheets, Calendar, Docs, Chat, Tasks, and 11 more Workspace APIs from the command line. Dynamic Discovery-based commands, helper shortcuts, schema introspection, and cross-service workflows.1.0.1HindsightLong-term memory for Claude Code backed by Hindsight. Hooks recall relevant memories before every prompt and retain the conversation after each turn via the installed `hindsight` CLI (external API). Includes a skill for manual recall/reflect/retain and bank management.1.0.0Hosted AgentsInfrastructure patterns for hosted background agents. Sandbox environments, image registry pattern, self-spawning agents, multiplayer support, warm pools, and multi-client integration (Slack, web, Chrome).1.0.5Mcp ServerComprehensive MCP (Model Context Protocol) server development skill. Build, configure, and manage MCP servers using Python (FastMCP) or TypeScript. Covers agent-centric design, tool creation, evaluation testing, production deployment, Claude Code integration, and plugin development.1.2.21Memory SystemsProduction memory architectures for LLM agents. Compares Mem0, Zep/Graphiti, Letta, Cognee, LangMem with benchmarks. Covers temporal knowledge graphs, memory consolidation, and retrieval strategies.1.0.6Multi Agent PatternsArchitecture patterns for multi-agent LLM systems. Supervisor/orchestrator, peer-to-peer/swarm, hierarchical patterns, context isolation, consensus mechanisms, and the telephone game solution.1.0.5Navigation DesignDesign information architecture, wayfinding systems, breadcrumbs, and navigation patterns for documentation and applications.1.0.11Nextjs DevelopmentComprehensive Next.js development skill covering App Router (13+/15/16), Server Components, Server Actions, Cache Components, data fetching patterns, and module architecture.2.0.24Ontology DesignDesign knowledge models with classes, properties, relationships, and taxonomies for structured data representation.1.0.11OsintOSINT intelligence skill — systematic person research from name/handle to scored dossier with psychoprofile, career map, and confidence grades. 55+ Apify actors, 7 search APIs, swarm mode.1.0.1Outcome OrientationFocus on measurable outcomes using OKRs, results-driven thinking, and outcome vs output distinction.1.0.11Persona DefinitionCreate detailed user personas with demographics, goals, pain points, behaviors, and empathy maps.1.0.11Persona MappingMap stakeholders and personas using Power-Interest matrices, RACI charts, and influence analysis.1.0.11Plugin DevEnd-to-end Claude Code plugin authoring toolkit. 8 skills covering the full lifecycle: ideation, research, architecture, hooks, composition, validation, evaluation, and documentation generation. Plus 4 runnable scripts: plugin scaffolder, structural validator, eval harness, and hook tester. Document any plugin by URL or local path with comprehensive README generation.1.2.0PrioritizationApply prioritization frameworks including RICE, MoSCoW, ICE scoring, and effort-impact matrices for decision-making.1.0.11Product ThinkingFive composable product-thinking skills: frame the real problem (JTBD, 5-whys), identify user needs (functional/emotional/social jobs), design value propositions (VPC, Kano), apply outcome-over-output thinking (North Star, leading/lagging metrics), and analyze trade-offs (cost-benefit, opportunity cost, reversibility, second-order effects).1.0.1Prompt EngineeringComprehensive prompt optimization system for LLMs. Design effective AI interactions, evaluate prompt quality, and perform iterative refinement for any LLM platform.1.1.16Python DevelopmentComprehensive Python development skill covering modern tooling (uv, ruff, mypy, pytest), best practices, coding standards, library architecture, functional patterns, async programming, MicroPython, and production-grade development workflows.1.1.25React DevelopmentBuild production-grade React applications and component libraries with shadcn/ui components, optimized hooks, client-side state patterns, and Bulletproof React architecture.1.1.21Research SynthesisMulti-source research coordination, evidence triangulation, competing-hypothesis analysis, and structured synthesis for knowledge-intensive tasks.1.0.0Risk ManagementApply risk assessment frameworks, mitigation strategies, risk registers, and monitoring practices.1.0.12Security EngineeringApplication security design, OWASP patterns, authentication architecture, secrets management, and threat-model-informed code review.1.0.0Skill FoundryFramework for creating Claude Code skills using philosophy-first design, evidence-based prompting, progressive disclosure, and anti-pattern prevention. 47 references, 25 scripts, 17 templates, 23 examples.2.2.2Social Media ContentLinkedIn posts, Twitter/X threads, short-form social content, platform-native tone adaptation, and engagement-optimised formatting.1.0.0Skillstack WorkflowsTwenty composable workflow playbooks that orchestrate existing SkillStack plugins for real multi-stage problems: plugin authoring, plugin updating, plugin/skill evaluation, API-to-production, security hardening, codebase onboarding, product stories, context engineering, design review, agent improvement, stakeholder storytelling, pitch sprints, complex debugging, AI agent building, strategic decisions, content platforms, user research, legacy rescue, LLM cost optimization, and skill authoring. Each workflow is a self-contained playbook with phase-by-phase guidance, gates/loops, and explicit references to the underlying SkillStack skills.2.2.1StorytellingStorytelling craft and application across fiction, business, data, speech, UX, and interactive narrative. Covers structural frameworks (3-act, Hero's Journey, Story Circle, Kishōtenketsu, Save the Cat, StoryBrand, Pixar Spine, Monroe's Motivated Sequence), narrative craft (character design, scene construction, dialogue, POV, pacing), domain applications, narrative theory (Propp, Polti, Booker, Tobias, Campbell), and anti-patterns. Progressive-disclosure structure with 12 domain references.1.0.1Systems ThinkingApply systems thinking principles including feedback loops, leverage points, and system dynamics to analyze complex problems.1.0.11Technical CopywritingFive composable skills for writing long-form technical content for an audience — articles, deep-dives, tutorials, newsletters, whitepapers, technical essays. Covers research before craft (audience profiling, source tiering, triangulation, citation), long-form structure (article templates, hook-promise-payoff contract, section transitions, length strategy), engaging craft (AIDA, PAS, Bencivenga's pyramid, Sugarman's slippery slide, Schwartz awareness levels, hooks, voice, concrete-over-abstract), long-form polish (pacing, scan-ability, the 30% cut, read-aloud test), and distribution craft (titles, dek/meta, social pull-quotes, channel framing). Grounded in proven copywriting techniques and evidence-based research workflow. Use when writing a technical blog post, deep-dive, tutorial, newsletter issue, whitepaper, or 1500-5000+ word technical essay. NOT for code documentation (use documentation-generator), UX microcopy (use ux-writing), short-form work writing like RFCs and emails (use communication), fiction or narrative craft (use storytelling), or CMS content models (use content-modelling).1.1.0Test Driven DevelopmentComprehensive Test-Driven Development skill implementing Red-Green-Refactor cycle across Python, TypeScript, JavaScript, and Emacs Lisp. Covers pytest, Vitest, Playwright, ERT, and Zod.1.1.18Testing FrameworkTest framework router and infrastructure setup across multiple languages and platforms. Selects focused modules for unit, integration, E2E, accessibility, mutation, fuzz, and CI/CD test integration.1.1.24Tool DesignDesign tools optimized for LLM agents rather than human developers. Consolidation principle, architectural reduction, tool description engineering, MCP naming, and the file system agent pattern.1.0.5Typescript DevelopmentComprehensive TypeScript development skill covering type system mastery, runtime validation (Zod, TypeBox, Valibot), framework integration (React 19, Next.js 16, NestJS, React Native), architecture patterns, security, tsconfig optimization, and testing strategies.1.1.21User Journey DesignDesign user journey maps with touchpoints, emotional states, pain points, and opportunities.1.0.11Ux WritingWrite effective microcopy, error messages, button labels, and interface text using UX writing principles.1.0.11Recommended Marketplaces
Acanilcancakir/claude-code
Acbc Codeacbcdev/claude-code
Agent Accordsdaniloc/agent-accords
Agent Harness Pluginscboone/agent-harness-plugins
Agentic Codingffalor/agentic-coding
Agent MarketplaceMrMYHuang/agent-marketplace
Journey fit
Primary fit
Build + agent-tooling is the canonical home for a 66-plugin library you pull into Claude sessions while designing and shipping software. Agent-tooling reflects installable skills for APIs, evaluation rubrics, and LLM project methodology—not a single launch or analytics surface.
How it compares
66-plugin curated skill marketplace, not a single Obsidian integration or token-optimizer bundle.
Common Questions / FAQ
Who is Skillstack for?
Solo and small-team developers building with Claude agents who need installable skills for APIs, evaluation, and LLM project architecture.
When should I use Skillstack?
Use it when you are implementing agent features, designing APIs, or benchmarking outputs and want pre-packaged methodologies instead of blank-slate prompting.
How do I add Skillstack to my agent?
Register the Skillstack marketplace in Claude Code, browse the 66 plugins by category, and install only the skills you need—such as agent-evaluation, api-design, or agent-project-development—for the current milestone.