anthropics/knowledge-work-plugins
148 skills269k installs3M starsGitHub
Install
npx skills add https://github.com/anthropics/knowledge-work-pluginsSkills in this repo
1Data Visualizationdata-visualization is a knowledge-work agent skill for solo builders who need to turn datasets into clear charts without guessing library APIs or chart grammar. It provides a structured selection guide that maps analytical intent—trends, rankings, part-to-whole composition, distributions, correlations, geography, and flows—to primary and alternative chart types, then backs those choices with Python implementation patterns across matplotlib, seaborn, and plotly. The skill emphasizes publication-quality output, design discipline, and accessibility including color theory, so figures work in blog posts, investor updates, internal dashboards, and ship-phase performance reviews. Use it whenever you are explaining metrics, comparing cohorts, or shipping narrative reports and want consistent, citable visuals instead of ad-hoc plotting snippets.8kinstalls2Code ReviewCode Review is an agent skill for solo and indie builders who want a consistent second pass on changes before merge. You invoke it with a PR URL, diff, or path when you need confidence that a patch is safe to ship—not another ad-hoc chat skim. The workflow walks security (including OWASP-oriented risks like injection and auth), performance (N+1, leaks, complexity), correctness (edge cases, errors, races), and readability, then returns actionable fixes with examples. It fits the Prism journey at Ship → Review as the primary shelf, and also helps during late Build when you are still iterating on backend or API layers. Intermediate complexity: you should be able to point the agent at real diffs and interpret severity. Pair it with your own CI and security tooling; this skill structures agent attention, not replaces automated scanners.5.4kinstalls3DocumentationDocumentation is an agent skill that helps solo and indie builders produce maintainable technical writing across common formats. It maps each doc type to a practical outline: READMEs that lead with purpose and a fast quick start; API pages with auth, errors, limits, and SDK examples; runbooks with prerequisites, procedure, rollback, and escalation; architecture docs with context, diagrams, decisions, and data flow; and onboarding guides that connect environment setup to people and systems. Five principles keep output usable—audience-first structure, useful information up front, concrete examples, freshness, and linking instead of duplicating. Invoke it when triggers like “write docs for” or “create a README” appear, or whenever you need structured help turning implementation into prose. It fits Build most often but the same skill supports operational runbooks during Operate and launch-facing README updates during Ship or Launch.5kinstalls4Task ManagementTask Management is a minimal, file-based commitment system for agents and humans sharing one repo. It always reads and writes TASKS.md in the current working directory, creating the standard four-section template if missing. On first task interaction it can copy dashboard.html from the plugin root so drag-and-drop boards stay synced with CLI edits. Tasks use checkbox lines with bold titles plus context, and completions are recorded with strike-through and dates. The skill is not a heavyweight PM suite—it is the glue for solo builders who want the agent to answer what is on my plate, add waiting-on items, or close work without spinning up Jira. It pairs with the broader knowledge-work productivity plugin for deeper setup while remaining usable journey-wide whenever you need honest task state in markdown.4.6kinstalls5Build DashboardBuild-dashboard is an agent skill that guides you through defining dashboard purpose, audience, key metrics, and slice dimensions, then assembling a professional interactive HTML report. Solo builders use it when they need an executive overview, a monitoring snapshot, or a deep-dive without standing up BI infrastructure. Data can come from live warehouse queries when connectors exist, uploaded CSV, pasted tables, or plausible sample data when you are prototyping layout first. The output is intentionally portable: charts, filters, and tables live in one file teammates can open offline. It fits indie SaaS founders who want investor-ready KPI views, ops leads who want filterable incident or usage panels, and builders who already have SQL results and need a polished narrative layer. Complexity is intermediate because you must clarify metrics semantics and data shape before the agent can embed JSON correctly.4.4kinstalls6Memory Managementmemory-management is a journey-wide agent skill for solo builders and tiny teams who already rely on Claude Code (or similar) daily and tire of re-explaining that “todd” is a person, “PSR” is a report, and “oracle” is a deal—not the vendor. It prescribes a deliberate architecture: keep a small, always-loaded CLAUDE.md with your highest-frequency people, terms, active projects, and preferences, while the expandable memory/ tree holds the full decoder ring—glossary, richer people and project notes, and company context. When you issue compressed instructions, the agent is supposed to expand them the way an insider would before planning or executing work. That pays off across Idea research notes, Validate scoping messages, Build implementation, Grow lifecycle emails, and Operate incident threads, anywhere internal vocabulary appears. The skill is meta—not a feature generator—but it compounds every other skill because plans and integrations reference real entities correctly. Setup is intermediate: you must curate and prune the hot cache so it stays within context budgets while remaining useful. Maintaining memory is ongoing editorial work, not a one-time install.4.4kinstalls7System Designsystem-design is an agent skill that walks solo builders through designing services, APIs, and architectures when prompts mention system design, architecture choices, data modeling, or service boundaries. It is not a code generator; it applies a five-part framework—requirements, high-level design, deep dive, scale and reliability, and explicit trade-offs—so decisions about storage, caching, queues, and failure modes are documented before coding. Indie founders use it to sanity-check whether a SaaS idea fits one Postgres instance or needs queues and replicas, and to produce contracts agents can implement later. It fits Validate when scoping an MVP, Build when splitting monolith boundaries, and Ship when hardening reliability narratives. Outputs emphasize assumptions, diagrammed data flow, and what to revisit as load grows—ideal when you wear architect and engineer hats and need citable structure instead of chat improvisation.4.2kinstalls8Knowledge SynthesisKnowledge Synthesis is an agent skill for the last mile of enterprise search: it takes raw results from Slack-style chat, email, cloud storage, and project trackers and produces a coherent, deduplicated answer with explicit attribution. Solo and indie builders who rely on agents for company or competitive research get a structured workflow for turning noisy multi-channel retrieval into decisions you can cite—API choices, ownership, and timeline alignment appear once, with sources listed by channel and date. The skill covers cross-source deduplication, confidence from freshness and authority, and effective summarization when hit counts are large. It is not a search connector itself; it assumes upstream search or MCP tools already returned fragments. Invoke when an agent must answer “what did we decide?” or “what is the current truth?” across systems without hallucinating consensus. Best paired with retrieval skills or internal search plugins before synthesis runs.3.7kinstalls9View PdfView PDF is an Anthropic knowledge-work plugin skill that gives agents a local PDF server and interactive viewer so solo builders can open contracts, forms, and papers and collaborate on markup in real time. It shines when the user wants to see the document, highlight terms, fill fields, stamp CONFIDENTIAL, or sign a specific page—not when they only need text. The skill draws a hard line: summarization, page quotes, and table extraction should go through the native Read tool because the viewer’s value is visual collaboration, not streaming bytes back to the model. Tools include listing local PDFs and display_pdf with optional starting page. For Prism’s audience, it is phase-specific agent tooling that also supports validate-time contract review and operate-time form handling, as long as interactivity is the goal.3.5kinstalls10Tech DebtTech Debt is an agent skill that systematically surfaces, classifies, and ranks technical debt so solo builders and small teams stop debating gut feelings and start working a backlog. It walks through six familiar debt types—from duplicated code and flaky tests to manual deploys and missing runbooks—and scores each item on impact, risk, and effort using an explicit formula. The output is not a vague “we should clean this up” memo but a prioritized remediation list with estimated effort, why each item matters to the business, and a phased plan meant to run beside normal feature delivery. Use it when maintenance feels opaque, regressions trace to old shortcuts, or you need to justify refactor time to yourself or stakeholders before the next sprint.3.5kinstalls11ArchitectureThe architecture skill guides an agent through creating Architecture Decision Records or evaluating system design proposals when you are choosing stacks, documenting trade-offs, or shaping a new component under constraints. Solo and indie builders use it to avoid reversible-by-chat decisions: Kafka versus SQS, microservices reviews, or notification-system design all land in a consistent markdown ADR your future self and collaborators can cite. It fits the build phase shelf as PM-grade documentation, but the same workflow applies while scoping in validate and while reviewing proposals before ship. Output follows a full ADR skeleton with options tables, pros and cons, and decision status so agents and humans align before heavy implementation spend.3.4kinstalls12Testing Strategytesting-strategy is an agent skill from Anthropic’s knowledge-work set that helps solo builders and small teams design pragmatic test strategies without defaulting to either zero tests or all-E2E suites. It encodes a classic pyramid—fast unit tests at the base, integration in the middle, selective end-to-end on top—and maps recommendations to component shapes such as HTTP APIs, ETL pipelines, frontends, and infra. The skill steers effort toward business-critical paths, failure modes, edge cases, security boundaries, and data integrity while naming what not to waste cycles on. Invoking it yields a concrete plan: what to test, which test type per area, coverage targets, and sample cases, plus identification of holes in current coverage. Use it while hardening a SaaS or API before launch, when refactoring backend logic in build, or when an agent is asked to “write tests for” a module without a plan.3.4kinstalls13Explore DataExplore-data is an agent skill for solo builders and small teams who need a disciplined first pass on unfamiliar datasets before SQL dives or ML work. It walks through accessing data—live warehouse via MCP, local files with inferred types, or user-guided queries when neither is connected—then builds a comprehensive profile of structure and quality. You invoke it when you inherit a new table, receive a CSV export, or need to spot null-heavy columns, odd distributions, and duplicate keys that would break downstream analytics. The skill keeps analysis honest by forcing shape-and-quality questions up front, which saves wasted build time on bad joins or meaningless KPIs. It pairs naturally with warehouse connectors documented in CONNECTORS.md and extends into Grow when you operationalize recurring analytics.3.1kinstalls14Incident Responseincident-response is an agent skill that guides solo builders and small teams through a full incident lifecycle: triage severity, assign incident commander and comms roles, publish status updates, and draft blameless postmortems after restore. It activates on phrases like “production is down,” “we have an incident,” or when an alert needs SEV grading—not for casual bug fixes in dev. The skill encodes a repeatable ritual so you communicate before you debug endlessly and document learnings while memory is fresh. It fits SaaS, APIs, and CLIs where downtime has user impact. Use it when you need structure under pressure; pair it with monitoring alerts and on-call rotations you already run. It is process knowledge packaged as slash commands, not an on-call pager replacement.3.1kinstalls15Analyzeanalyze is an agent skill that answers data questions in plain language for solo builders who have a warehouse or database MCP hooked up. It starts by classifying whether you need a single KPI, a multidimensional investigation (for example what drove a conversion drop), or a formal report with caveats and recommendations. Then it maps your question to schema, writes SQL, executes through connectors, and formats results appropriately. Use it in Grow when reviewing weekly signups, in Validate when sanity-checking pricing or funnel assumptions from early data, and in Operate when correlating errors with usage spikes. It does not replace a governed BI stack—it accelerates ad-hoc analyst work inside your coding agent. Intermediate complexity: you should understand basic SQL and what tables mean.3.1kinstalls16Search StrategySearch Strategy is the orchestration brain behind Anthropics knowledge-work plugins: it takes a plain English question such as what was decided about an API migration timeline and explodes it into targeted parallel queries across chat, knowledge bases, and project trackers, each with syntax tuned to that source. Solo builders and tiny teams reuse the same pattern when their “enterprise” is scattered across Slack, Notion, Linear, and email—without writing bespoke search prompts per tool. The skill documents query-type classification first, then decomposition, relevance ranking, and synthesis so the agent does not treat search as a single generic web call. It is marked user-invocable false because it is procedural glue agents invoke while answering, not a button users mash directly. Use it whenever accuracy depends on finding conclusions and recent status signals, not just keyword luck. CONNECTORS.md remains the source of truth for which tools are wired so strategies stay honest about coverage.3.1kinstalls17DebugDebug is an agent skill that turns ad-hoc firefighting into a structured debugging session you invoke with /debug plus an error message, stack trace, or problem description. It is built for solo and indie builders who ship with Claude Code, Cursor, Codex, or similar agents and need a repeatable way to move from “something broke” to a verified fix without skipping reproduction or scope. The workflow walks you through reproducing the failure, isolating the component or change that introduced it, diagnosing the root cause from logs and recent changes, and applying a fix with evidence. Triggers align with real builder pain: errors after deploy, behavior that works in staging but not production, and divergent output when the cause is not obvious. Use it during backend work, after release, or while operating live systems whenever systematic isolation beats guessing in chat.3.1kinstalls18Content Creationcontent-creation is a marketing writing skill that gives coding agents structured templates for blogs, social posts, email newsletters, landing pages, press releases, and case studies. Solo founders who ship product themselves often stall on distribution because every channel expects a different shape of copy; this skill standardizes hooks, body sections, conclusions, and calls to action while nudging SEO basics such as headline length and meta descriptions. It is not a CMS integration or analytics suite—it is procedural guidance the model follows when you ask for channel-specific formatting, headline options, or keyword-aware drafts. Use it while growing an audience after launch, while preparing launch narratives and landing copy, or during validation when you need a credible landing page narrative before the product is finished. The knowledge-work-plugins packaging assumes Claude-style agents; outputs are drafts for human edit, not automated publishing. Pair with separate SEO or distribution skills when you need technical site audits or paid channel playbooks.3kinstalls19Source ManagementSource Management is an Anthropic knowledge-work plugin skill that keeps enterprise search honest about what MCP connectors are actually live. Solo builders and small teams running Claude with workplace integrations use it so the agent does not hallucinate access to chat, email, drive, project trackers, CRM, or knowledge bases. The procedure explains how to infer connected sources from available MCP tool prefixes, what each source can query, and how to coach users to connect missing systems in MCP settings. It also covers ordering sources by priority and staying aware of rate limits when blending results. Because it is user-invocable false, it shapes agent behavior during knowledge tasks rather than acting as a standalone slash command. Pair it with CONNECTORS.md when placeholders or tool names are unclear. Best when you are standing up or extending an internal copilot that must search across many SaaS silos without brittle hard-coding.3kinstalls20Deploy ChecklistDeploy Checklist is a Claude knowledge-work agent skill that turns an imminent release into a repeatable shipping gate. Solo builders and small teams use it when they are about to ship, especially when the change includes database migrations, feature flags, or anything that needs documented rollback criteria before production. The skill emits a dated markdown checklist covering CI and review readiness, staging verification, smoke tests, canary or full production rollout, short-window observability, and post-deploy stakeholder updates. Because it accepts a service or release name as an argument, you can adapt the template in conversation without maintaining a separate runbook doc. It fits the Ship phase as a launch-prep artifact: you still own your thresholds and tooling, but the agent standardizes what must be true before and after the button press.3kinstalls21Create Vizcreate-viz is an Anthropic knowledge-work agent skill that standardizes how coding agents produce publication-quality visualizations in Python. Solo builders use it when SQL or spreadsheet results need to become trustworthy charts for investor updates, growth reviews, or product postmortems—not ad-hoc matplotlib snippets that mis-scale axes or hide uncertainty. The workflow starts by pinning the data source (warehouse query, upload, or prior analysis), the chart intent, and the audience; then it loads pandas DataFrames, cleans types, selects an appropriate visual encoding, and renders with design defaults suited to reports and presentations. It explicitly supports optional warehouse connectors when configured via CONNECTORS.md. Intermediate complexity: you should understand basic DataFrames and when a line chart beats a bar chart, but the skill carries layout and accuracy conventions so one-person teams do not ship misleading graphics.2.9kinstalls22StandupStandup is a knowledge-work plugin skill that generates a daily standup update by formatting your work into yesterday, today, and blockers—and optionally aggregating recent commits, pull requests, ticket changes, chat context, and CI/CD signals when connectors are configured. Solo and indie builders use it before async standups, investor updates, or co-founder syncs when they have fragmentary notes but need a tight narrative. It always works in standalone mode from a short brain dump; connected tools reduce manual copy-paste from GitHub, Jira-style trackers, and chat. Prism places it on Operate → iterate as the habitual check-in shelf while noting real data often spans Build and Ship. Pair with your connector docs so the agent knows which integrations are live.2.9kinstalls23SearchSearch is an agent skill command for solo and indie builders who already use MCP to wire chat, email, cloud storage, project trackers, CRMs, and knowledge bases into Claude Code, Cursor, or similar agents. It is meant for moments when you remember a decision, doc, or thread existed somewhere but not which tool holds it—phrases like “find that doc about…”, “what did we decide on…”, or “where was the conversation about…”. The workflow checks which connectors are live, parses intent and time scope, fires parallel targeted searches per source, deduplicates overlaps, and returns a short synthesis with concrete citations rather than a raw dump. When nothing is connected, it tells you how to fix MCP settings instead of failing silently. Because recall spans research, build, ship prep, and day-two operations, treat it as journey-wide glue you invoke whenever scattered organizational memory blocks progress.2.8kinstalls24UpdateUpdate is a productivity command skill for solo and indie builders who treat TASKS.md and a memory/ folder as the source of truth for agent-assisted work. In default mode it loads current state, pulls open assignments from connected project trackers via MCP when available, and falls back to GitHub Issues with `gh issue list --assignee=@me`, then compares each external item to TASKS.md so you can add missing work instead of duplicating tickets. It also supports triaging stale or overdue lines and filling gaps when people or projects appear in context but not in memory. Comprehensive mode widens the net to chat, email, calendar, and documents to catch todos that never became tasks and to propose new memory entries. The skill is procedural: it assumes you already wired connectors and may point you to start if the productivity scaffold is absent. Use it whenever your external board moved faster than your repo-local plan.2.8kinstalls25Competitive Intelligencecompetitive-intelligence is an agent skill that turns open-web research into an interactive HTML battlecard for solo founders and small teams selling B2B or competitive SaaS. It works standalone: the agent deep-dives each competitor’s product, pricing, and positioning, summarizes what they shipped in roughly the last ninety days, contrasts your recent releases, and packages differentiation, talk tracks, and landmine questions into clickable cards plus a comparison matrix. Use it when a prospect names a rival, when you refresh positioning before validation, or when launch and growth motions need crisp “why us” artifacts without a full PMM hire. Output is a single HTML artifact meant for live deals—not a slide deck template. It does not replace legal review of claims or proprietary win/loss data you have not supplied.2.8kinstalls26StartStart is the entry workflow for Anthropic’s knowledge-work productivity plugin, aimed at solo builders who want one place for tasks, shorthand todos, and working memory inside their agent session. Invoke it when you set up the plugin for the first time, need to bootstrap CLAUDE.md and the memory/ tree from an existing task list, or want nicknames and project codenames decoded into structured tasks. The skill walks through detecting what already exists, creating missing TASKS.md and dashboard.html in the current working directory, and orienting you to open dashboard.html from your file browser—explicitly avoiding xdg-open in VM Cowork environments. After the dashboard step, a fresh install continues into memory bootstrap so agents retain context across sessions. It pairs with the broader task-management skill referenced in the instructions and CONNECTORS.md for tool wiring. Complexity is beginner-friendly setup with ongoing multi-phase value once memory is populated.2.8kinstalls27Legal Risk AssessmentLegal Risk Assessment is an agent skill that walks you through classifying legal issues using a structured severity-by-likelihood framework, then documents when escalation is warranted. It is built for in-house-style workflows but fits solo and indie builders who negotiate SaaS terms, vendor MSAs, affiliate deals, or client SOWs without a full legal bench. Invoke it when you are evaluating contract clauses, estimating financial or reputational exposure, or sorting a backlog of open legal questions by priority. The skill emphasizes consistent labels across deals so you do not treat every redline as catastrophic or ignore material liability. Outputs are assessment-ready narratives for qualified counsel to review—not autonomous legal opinions. Pair it with your own playbook thresholds so negligible/low items stay in normal ops while moderate-and-above items get explicit review gates before signature or launch.2.7kinstalls28Product BrainstormingProduct Brainstorming installs a seasoned PM or design-lead voice into your agent session so you can think out loud before anything becomes a PRD. Solo founders often skip structured discovery and jump straight to features; this skill keeps you in problem space long enough to map who hurts, what they do today, and what breaks if you do nothing. It offers distinct brainstorming modes—problem exploration when the space is undefined, ideation when you need volume, and stress-testing when you already lean toward one bet—so the agent adapts instead of dumping generic lists. The skill is journey-wide because the same sparring partner helps at Idea when sizing an opportunity, at Validate when scoping a prototype, at Build when reframing a stuck feature, and at Launch when positioning needs a sanity check. Complexity sits at intermediate: you bring context and willingness to be challenged. Outputs are clarity and direction, not files; pair it with planning or writing skills once you converge.2.6kinstalls29Sql QueriesSQL Queries is an agent skill that helps solo and indie builders write correct, readable, and performant SQL across major analytical warehouses and PostgreSQL-style engines. Invoke it when you are drafting new queries, refactoring slow reports, translating logic between dialects, or composing multi-step analytics with CTEs, window functions, and aggregations. The bundled reference foregrounds PostgreSQL semantics—interval math, truncation, JSON path access, and array helpers—while the stated scope explicitly includes Snowflake, BigQuery, Databricks, and similar platforms so you do not mix incompatible functions. It fits builders shipping SaaS or API products who wear the data hat without a dedicated analytics team: wiring Supabase or Neon for product analytics, debugging a dashboard in Metabase or Hex, or optimizing a nightly job before costs spike. Use it during grow-phase metric work and whenever build or operate tasks need dependable warehouse SQL rather than copy-paste from chat.2.5kinstalls30Statistical AnalysisStatistical Analysis is an agent skill that walks solo builders through descriptive statistics, trend checks, anomaly detection, hypothesis testing, and honest interpretation of results. It starts with practical methodology: when symmetric data warrants the mean, when skewed revenue or usage metrics demand the median, and why reporting both center measures prevents false confidence in a single number. Spread guidance covers standard deviation for roughly normal metrics, IQR for heavy tails, and coefficient of variation for comparing volatility across scales with different units. Invoke the skill while exploring distributions, testing whether a launch lift is significant, correlating engagement with retention, or flagging outliers before they distort dashboards. It complements raw SQL or spreadsheet exports—you still bring the dataset, but the agent applies consistent statistical framing so indie operators do not overclaim from thin samples or ignore skew. Most valuable in grow analytics, it also supports validate-phase experiment readouts and build-phase ML or scoring features that need sane baselines.2.5kinstalls31Financial Statementsfinancial-statements is a knowledge-work agent skill that structures how you generate P&L-centric reporting with comparisons and variance narrative for monthly, quarterly, or annual periods. It targets solo founders and indie operators who must produce credible financial summaries for themselves, advisors, or investors without a full finance team—while still routing final numbers through qualified professionals. The workflow centers on income statement generation with hooks to balance sheet and cash flow formats, GAAP-oriented presentation notes, and common period-end adjustments as reference. Connectors may pull actuals from linked tools per CONNECTORS.md. Use when closing the books, comparing actuals to budget, or preparing a leadership packet—not when you need tax filing automation or personalized investment advice. Argument hint documents frequency and period explicitly so agents do not guess the reporting window.2.4kinstalls32Design CritiqueDesign Critique is an agent skill that turns a vague “what do you think of this screen?” into a repeatable usability and visual-design review. Solo and indie builders paste a Figma URL, attach a screenshot, or describe a layout, then receive feedback organized around a clear framework: two-second first impression, goal completion and navigation, visual hierarchy and emphasis, cross-screen consistency, and final polish. You can narrow the lens to mobile or a specific flow such as onboarding. It is aimed at anyone shipping web or mobile product UI without a dedicated design partner, and fits whenever you are exploring concepts, refining a prototype, or doing a last pass before release. The output is actionable critique you can feed back into your design tool or implementation plan, reducing rework from hierarchy mistakes and unclear CTAs.2.2kinstalls33User ResearchUser Research is an agent skill that helps solo and indie builders plan, run, and synthesize qualitative and quantitative studies when they need to understand real users—not guess from vibes. Trigger it with phrases like “user research plan,” “interview guide,” “usability test,” “survey design,” or “research questions,” or whenever you are deciding how to learn what people need, whether they can use a flow, or how to rank discoveries. The skill encodes a practical methods matrix so you pick interviews, usability tests, surveys, card sorting, diary studies, or A/B tests for the right question and timeline. It outputs structured interview scripts and analysis patterns such as affinity mapping, impact/effort prioritization, journey maps, and jobs-to-be-done framing. Use it before committing to an MVP, while validating a prototype, and again after launch when behavior shifts. It is editorial guidance packaged for agent workflows, not a recruiting panel or analytics warehouse.2.1kinstalls34Ux CopyUX Copy is an agent skill for writing and reviewing interface language—microcopy, error messages, empty states, CTAs, confirmations, and onboarding—in product flows built by solo developers who do not have a dedicated content designer. Invoke it when you are naming a button, wording a destructive dialog, or polishing first-run text and want consistency with a small set of principles: clarity, brevity, consistent terminology, usefulness, and a human tone. The skill expects you to supply screen context, user intent, emotional state, desired tone, and any character or platform limits; it returns copy patterns such as verb-first CTAs and three-part errors. It fits Claude Code and Cursor workflows where the same agent that implements UI also drafts strings. Use during build for new screens, during ship for pre-release copy review, and at launch when onboarding and paywall wording affects conversion—without replacing brand voice guidelines you already own.2kinstalls35Slack MessagingSlack Messaging is an agent skill for solo builders and small teams who relay updates through Slack from Claude Code, Cursor, or similar agents. Slack clients do not render GitHub-flavored Markdown; they use mrkdwn, so pasted plans and status reports often show raw asterisks or dead links. The skill documents the formatting table—bold, italic, strikethrough, inline and fenced code, quotes, `<url|label>` links, and `<@U…>` / `<#C…>` mentions—and lists common mistakes such as double-asterisk bold and header syntax that Slack ignores. Apply it whenever you draft or send Slack messages or canvases so automation outputs look professional in channels without manual cleanup. It is procedural knowledge, not a separate MCP server; pair it with your Slack integration tools at send time.2kinstalls36Design Systemdesign-system is a knowledge-work agent skill for solo builders and tiny teams who cannot afford a dedicated design-ops role but still want systematic UI. Invoke it to run a full consistency audit—surfacing naming drift and hardcoded values across components—or to produce documentation for a single component’s variants, states, sizes, interactions, and accessibility expectations, or to sketch a new pattern that fits existing tokens and components. The SKILL.md structures work around atomic tokens, reusable components, and composite patterns such as forms and navigation, with explicit argument hints for audit, document, and extend paths. It is most natural while building frontend surfaces in React, Vue, or similar stacks, but audits also help in Ship when you are hardening UI before launch. Compatible with Claude-style plugins and other agents that read CONNECTORS.md for tool wiring. The outcome is clearer shared language between you and your coding agent so new screens do not invent one-off colors, spacing, or interaction states.2kinstalls37Research SynthesisResearch Synthesis is an agent skill for distilling messy user research into patterns solo builders can act on. When you have interview transcripts, pasted survey results, usability test notes, support tickets, or NPS and app store feedback, the skill produces a Nature-style research report: executive summary, key themes with prevalence counts and quoted evidence, and a table mapping insights to opportunities and impact. It is aimed at indie founders and small teams who run lightweight research without a dedicated UX researcher—often right after customer interviews or before rewriting a landing page. Use it whenever qualitative data exists but nobody has time to code themes manually. The workflow references the broader user-research skill for interview guides and frameworks while keeping this invocation focused on synthesis and prioritized next steps. Outputs are markdown-ready for specs, Notion, or handoff to planning skills.1.9kinstalls38Interview PrepInterview Prep is an agent skill for founders and indie operators who are about to run real interviews and want the same rigor larger teams use without an HR department. It walks you through defining four to six competencies, building a question bank per competency with behavioral and situational prompts, and attaching a consistent 1–4 rubric plus a debrief format so every candidate sees the same bar. Trigger it when you need an interview plan, role-specific questions, or a scorecard before talking to applicants. The output is a complete kit—panel roles, questions by competency, scoring guidance, and debrief structure—so agents and humans can run interviews consistently and document evidence for decisions.1.9kinstalls39Review Contractreview-contract is a checker-oriented workflow from Anthropic's knowledge-work plugins that helps solo builders and tiny teams stress-test SaaS vendor agreements, customer MSAs, and partnership drafts against a written negotiation playbook instead of skimming in chat. The agent ingests the contract from upload, CLM link, or paste, gathers deal context, walks clauses in order, marks departures from standard positions, and outputs redline language plus prioritized fallbacks for business—not legal—judgment. It fits pre-signature validation when you are choosing infra stacks, analytics tools, or first enterprise customers, and again in Operate when renewals re-open liability and auto-renew traps. CONNECTORS.md governs which document systems are available; nothing here replaces licensed attorneys.1.8kinstalls40Accessibility Reviewaccessibility-review is an agent skill that walks solo builders through a WCAG 2.1 AA audit on a design URL, page URL, or UI description. It is aimed at indie developers and designers who ship their own interfaces and cannot afford an accessibility surprise after implementation. Use it when you are reviewing color contrast, keyboard navigation, touch target size, screen reader behavior, or form labeling—not as a substitute for full legal compliance programs, but as a structured pre-handoff gate. The skill embeds criterion-level reminders (1.1.1 alt text, 1.4.3 contrast, 2.1.1 keyboard, 2.5.5 targets, 4.1.2 name/role/value) so the agent’s output references real WCAG expectations instead of vague “make it accessible” advice. That matters for Prism’s audience because accessibility fixes are cheapest before code lands.1.8kinstalls41Daily BriefingDaily Briefing is an agent skill that turns whatever you know about today’s sales calendar and pipeline into a short, prioritized action plan. Solo builders and small teams selling their own product can use it without plugins by stating meetings, hot deals, and top priorities; the skill structures output for a quick morning read rather than another long doc. When you connect calendar, CRM, and email, the same workflow can pull meetings with attendees, pipeline alerts, unread messages from key accounts, and overnight signals so prep scales beyond manual recap. It is aimed at revenue-facing operators who need focus before deep work, not at generic project management. Complexity stays beginner-friendly because the default path is conversational input plus structured output. Pair it with your existing CRM or inbox workflows rather than treating it as a replacement for those systems.1.8kinstalls42Brand Voice Enforcementbrand-voice-enforcement is an agent skill from Anthropic’s knowledge-work plugins that turns bland GTM copy into voice-aligned, evidence-backed messaging. Solo founders use it when they have draft emails, landing blurbs, or social posts that sound like every other AI paragraph but lack the confident, data-driven tone their guidelines demand. The skill walks agents through concrete before/after examples — such as cold outreach — and labels what changed at the voice dimension so you can apply the same transforms to your real product names, ICP, and metrics. It is not a substitute for a full brand book; you still supply positioning, tone rules, and approved stats. It shines in the Grow and Launch phases whenever you are scaling outbound, refreshing lifecycle emails, or tightening copy before distribution pushes, and it pairs well with human review for compliance and factual accuracy.1.8kinstalls43Write SpecWrite Spec is an agent skill for solo builders who need a real feature specification instead of a one-line ticket. You start from anything from “SSO support” to “enterprise customers want centralized auth” or “something about onboarding drop-off,” and the workflow walks you through the user problem, segments, how you will measure success, constraints, and prior art without dumping a giant questionnaire at once. The skill is built for knowledge-work and plugin-style agent setups where turning requests into a structured PRD is the gate before implementation. It fits validate and early build moments when scope, non-goals, and phased delivery need to be explicit for you or a coding agent. Use it when the idea is directionally right but the document is not—so downstream planning and implementation have testable acceptance criteria rather than assumed intent.1.8kinstalls44Draft Outreachdraft-outreach is an agent skill for indie and solo founders who need cold email and LinkedIn-style outreach that sounds like it was written for one person, not a mail-merge. It mandates a research-first sequence: web search by default, augmented with enrichment and CRM connectors when you connect them, then a draft tailored to what you learned about the person or company. Without connectors it still works—you get polished message text to paste into your client. With connectors it can pull verified emails, prior relationship notes, and optionally stage a draft in your inbox. It fits the growth phase when you are doing founder-led sales, partnership pings, or investor outreach and cannot afford spray-and-pray copy. The skill is procedural workflow, not sending automation—you remain responsible for review and send.1.8kinstalls45Account ResearchAccount Research is an agent skill for solo and indie builders who sell services, SaaS, or consulting and need a fast briefing on a company or person before email, DM, or discovery calls. It always runs on web search to surface what the business does, size and industry context, recent news, hiring signals, leadership names, and what they sell. When you connect enrichment or your CRM, the same flow adds verified contact paths, technology footprint, and relationship history so outreach is specific rather than generic. Use it when a trigger names a prospect or account and you want actionable intel in one pass. It matters because weak pre-call research wastes cycles and burns trust; a consistent research ritual keeps your pipeline work tied to real signals instead of guesses.1.8kinstalls46Validate Datavalidate-data is an agent skill from Anthropic’s knowledge-work plugins that acts as a pre-flight reviewer for any analysis you are about to share. Solo builders wearing analyst and PM hats can point it at a report snippet, SQL output, notebook, or chart description and get a disciplined review of framing, dataset choice, time ranges, population definitions, metric consistency, and whether conclusions actually ride on the evidence. It fits the Grow phase when analytics drive roadmap calls, but it is equally valuable in Validate when a landing-page metric story must hold up, or in Ship when you sanity-check launch KPIs. The /validate-data command pattern keeps reviews repeatable instead of one-off chat nagging. CONNECTORS.md is referenced when tool wiring is unclear—plan on having analysis context in-thread or as files the agent can read.1.7kinstalls47ReconciliationReconciliation is a finance operations skill from Anthropic’s knowledge-work plugins that teaches agents how to compare general ledger control balances to subledgers, bank statements, and intercompany detail without substituting for professional judgment. Solo founders with revenue often still need monthly bank recs and AR/AP tie-outs even when they are not running a finance team; the skill walks through period-end pulls, variance investigation, aging of reconciling items, and when to escalate. It fits the Operate phase when books must stay audit-ready, not the Idea or Validate shelves where you are still proving demand. Pair it with exports from your accounting stack and human sign-off before filings. The argument-hint pattern supports scoping by account and period for repeatable close rituals.1.7kinstalls48Risk Assessmentrisk-assessment is a knowledge-work agent skill that walks solo builders and small teams through operational risk identification, scoring, and mitigation planning. It applies a standard likelihood-by-impact matrix, six category lenses, and a register format so decisions about projects, vendors, processes, or launches are documented with owners and status. Invoke when you need to answer what could go wrong, build a risk register, or stress-test a plan before spend or reputation is on the line. The methodology is domain-agnostic: the same structure works when validating a product bet, signing a vendor, or changing production process. It complements specs and PRDs by making tradeoffs explicit. Outputs are narrative registers suitable for stakeholders, not automated ticketing—your agent drafts and you assign real owners outside the chat.1.7kinstalls49Seo AuditSEO Audit is an agent skill that runs comprehensive search audits for solo builders and small marketing stacks who need more than a generic checklist. You can scope a full site review, focused keyword research, content gap analysis against competitors, a technical crawlability and structured-data pass, or a head-to-head competitor SEO comparison. The skill expects a URL or domain (or topic in keyword mode) and an audit type, then walks connected tools where available so output reflects real crawl and rank context rather than guesses. It is built for the Launch phase when a product site, landing page, or content property must earn traffic, but the same runs help Validate when you compare competitor positioning and Grow when you refresh content and close gaps. Finished work is framed as an executable prioritized plan—quick wins you can ship this week plus strategic investments—so a single operator can hand results to implementation without reformatting. Pair it with your site repo or CMS workflow in Claude Code or similar agents when assess SEO health, find keywords competitors own, or need a technical plus content gap pass in one session.1.7kinstalls50Design HandoffDesign Handoff is an agent skill that generates comprehensive developer handoff documentation from a Figma URL, screenshot, or written design description. It is aimed at solo and indie builders who ship UI with Claude Code, Cursor, or similar agents and need a single spec sheet instead of scattered comments. Use it when a design is ready for engineering and you want explicit layout measurements, token references, responsive behavior, interaction and animation details, content limits, edge cases, and accessibility requirements in one pass. The skill follows a structured sections model (visual, interaction, content, edge cases, accessibility) and can integrate with connected design tools per the plugin’s connector docs. It reduces rework from ambiguous mocks and gives implementers a checklist-aligned brief that pairs well with frontend implementation skills and code review.1.7kinstalls51Slack SearchSlack Search is an agent skill for solo and indie builders who run Claude Code, Cursor, or similar agents against Slack workspaces. It documents when to call each search primitive, how to respect user consent for private channels and DMs, and a repeatable query strategy: start broad, refine with operators, switch between conceptual questions and keyword queries, and iterate with related terms. The skill is for gathering context before answering “what did we decide?” questions—not for replacing Slack’s UI. It matters in Prism’s journey because comms context spans shipping incidents, launch coordination, and day-to-day ops; tagging it under Build → integrations keeps MCP-style Slack tooling discoverable while secondary phases cover support and lifecycle lookups.1.7kinstalls52Audit SupportAudit Support is an agent skill that walks you through Sarbanes-Oxley Section 404 internal control over financial reporting (ICFR) testing. It is aimed at solo builders and small teams who wear finance hats at funded or regulated SaaS companies and need structured help generating testing workpapers, choosing samples, documenting results, and classifying control deficiencies before internal or external audits. The skill encodes a repeatable methodology from scoping significant accounts through design and operating effectiveness tests, without substituting for auditors or legal counsel. Use it when you must translate control narratives into testable evidence, not when you only need generic security scanning. Outputs are professional-grade documentation patterns that still require review by qualified financial professionals.1.7kinstalls53Data Context Extractordata-context-extractor is a meta-skill from Anthropics knowledge-work-plugins that turns analyst tribal knowledge into a durable agent skill for your data warehouse. In Bootstrap mode it triggers on phrases like creating a data context skill or setting up analysis for your warehouse, then walks through database connection discovery, schema exploration, and targeted questions before generating an initial skill with reference files. Iteration mode loads an existing package when you need more context about a domain, metrics, or tables, and merges new answers into focused reference documents rather than bloating a single prompt. Solo founders with a small data team use it to stop re-explaining the same revenue definitions, event names, and join keys on every analysis request. The skill is procedural and interview-driven—success depends on accurate inputs from people who know the warehouse. Output is a maintained skill tree Claude can invoke for grounded SQL and narrative analysis aligned to your company semantics.1.7kinstalls54Customer ResearchCustomer Research is an agent skill for solo builders and tiny teams who handle their own support, success, and pre-sales questions without a dedicated research desk. You invoke it when a customer asks something you must verify, when you need to know if an issue was already reported, when you want prior messaging for a named account, or when you need topic background before drafting a reply. The skill walks through parsing the request type, narrowing what “done” looks like, querying every connected source, and returning a synthesized answer with source attribution and confidence—not a single-chat guess. It fits naturally in Grow support work, but the same pattern helps during Validate when scoping from real user questions and during Operate when triaging recurring production issues. Pair it with your CRM, docs, and ticket connectors documented in the plugin’s CONNECTORS file.1.7kinstalls55Process OptimizationProcess Optimization is a journey-wide agent skill for solo builders and tiny teams who describe a workflow as slow, bloated, or bottlenecked. It walks through mapping the current state step by step—including owners, duration, manual work, and approvals—then classifies waste across waiting, rework, handoffs, over-processing, and automation gaps. The future-state design pushes elimination, automation, fewer handoffs, parallel work, and lightweight checkpoints instead of approval gates. Closing metrics cover cycle time, errors, cost, and satisfaction. Invoke it whenever process improvement language appears, not only after launch: the same framework applies to validation rituals, release trains, support triage, and growth experiments that accreted too many steps.1.7kinstalls56Competitive BriefCompetitive Brief is an agent skill from Anthropic’s knowledge-work plugins for founders and solo marketers who need a disciplined competitor pass without spinning up a consulting deck. You supply one or more competitor names and optionally your product context, positioning, and differentiators; you can narrow focus to messaging, product features, content strategy, recent news, public pricing, or market presence—or let the skill cover all. It researches each competitor via web search and returns a structured comparison highlighting content gaps, messaging angles competitors have not claimed, opportunities, and threats—usable for sales battlecards or a quick read when a rival ships. Canonical placement is early journey competitor research, but the same artifact informs validate scope decisions and launch positioning refreshes. Intermediate complexity: quality improves when you provide honest differentiation inputs rather than only names. Connectors and tool availability are documented separately in CONNECTORS.md when placeholders appear.1.7kinstalls57Ticket TriageTicket-triage is an agent skill for solo builders and tiny teams who cannot afford a dedicated support desk but still need consistent handling when messages pile up. It reads a customer report or internal issue description, extracts what is actually broken, who is affected, and how urgent the situation is, then maps that to a P1–P4 priority frame and a routing recommendation. The workflow is built around operational clarity: separate symptoms from the core problem, surface production and multi-user risk, and note frustration or escalation risk so your reply tone matches the situation. Output is a structured triage write-up you can paste into Zendesk, Linear, GitHub Issues, or email—not a black-box auto-close. Use it the moment a new ticket arrives, before you ping engineering or refund someone. It pairs naturally with connector docs in the plugin when you later automate lookups; on its own it is judgment and formatting for fast, repeatable first responses.1.6kinstalls58Variance AnalysisVariance-analysis is an agent skill that helps solo founders and small finance owners explain why numbers moved without building a full FP&A team. It walks through standard decomposition—especially price and volume effects—with formulas you can sanity-check, plus guidance on when to separate mix, how to set materiality thresholds, and how to turn math into short narratives executives actually read. The skill explicitly disclaims financial advice: outputs are workflow assistance meant for review by a qualified professional before external reporting. Use it when you are closing a month, comparing forecast to actual, or prepping board-style commentary on revenue, expenses, or unit economics. It is methodology-heavy rather than a live ledger integration, so you bring the figures (CSV, spreadsheet exports, or pasted tables) and the skill structures the analysis. Intermediate complexity reflects comfort with basic accounting vocabulary and tolerance for checking arithmetic.1.6kinstalls59Compliance CheckCompliance-check is an agent skill from the knowledge-work plugins that walks solo founders and small teams through a compliance assessment of a proposed action—new product features, referral programs, biometric login, cross-border data processing, or testimonial-heavy ads. You invoke it with /compliance-check and a plain-language description of the initiative; the agent returns a structured markdown brief covering applicable regulations, approval paths, jurisdictional nuances, and risk areas. It is designed as a pre-ship guardrail when GDPR, HIPAA-flavored concerns, financial promotions, or consumer protection rules might apply, while repeatedly reminding you that outputs are not legal advice and must be verified with qualified counsel. Prism catalogs it for builders who wear product, marketing, and ops hats and need a repeatable checklist before engineering sprints or campaigns go live.1.6kinstalls60Meeting BriefingMeeting Briefing is an agent skill that acts as a meeting preparation assistant for in-house legal teams. It walks through identifying the meeting type and participants, assessing what preparation each scenario requires via a meeting-type table, gathering context from connected sources when available, and outputting structured briefings suited to contract negotiations, board meetings, compliance reviews, and other legally sensitive sessions. Solo builders wearing legal-adjacent hats on small deals—vendor terms, fundraising governance, policy reviews—can use the same structure even without a full legal org: you get consistent sections for background, risks, open questions, and desired outcomes instead of ad-hoc notes. The skill stresses human review before reliance and does not substitute for counsel. Use it when calendar-driven work needs a repeatable prep ritual and when meetings spawn follow-on tasks that must not get lost in chat logs.1.6kinstalls61Call PrepCall Prep is an agent skill that gets a seller ready for a specific meeting in minutes. In standalone mode you supply the company, meeting type, and attendees; the agent performs company and web research—recent news, funding, leadership shifts—and returns a concise brief with agenda and questions. When you connect CRM, email, chat, or call transcripts, the same flow enriches with account history, open threads, internal chatter, and prior commitments so you are not re-asking answered questions. Solo founders doing their own sales benefit because the skill encodes a repeatable prep ritual instead of scrambling through tabs before every Zoom. It targets growth-stage conversations rather than one-off launch announcements: discovery, demos, renewals, and expansion. It does not replace CRM hygiene or recorded call analytics; it aggregates what you already have plus public signals into one agent-readable packet.1.6kinstalls62Campaign PlanCampaign Plan is an agent skill that interviews you for campaign goal, audience, timeline, and budget, then drafts a full marketing brief you can execute as a solo founder or tiny team. It is aimed at builders who have something to ship but lack a marketing ops bench—product launches, signup pushes, awareness sprints, and win-back motions. Output spans positioning and messaging, channel mix, a dependency-aware content calendar, and measurable success criteria so you are not guessing what to post each week. It aligns with Prism’s launch and grow journey: use it when you are moving from validated offer to coordinated distribution, not when you only need a single tweet or ad variant. Connectors referenced in the plugin may extend reach, but the core value is a structured, citable plan artifact.1.6kinstalls63Performance ReviewPerformance Review is an Anthropic knowledge-work agent skill that structures review season paperwork: self-assessments, manager reviews for named direct reports, and calibration prep. Solo builders with one or two hires—or indie founders who suddenly manage contractors—use it when review season kicks off and blank Google Docs stall progress. The skill asks for a mode if you do not specify one, then emits markdown templates with accomplishments, goals tables, and calibration-oriented framing rather than generic praise. It connects conceptually to CONNECTORS.md when placeholders or integrated tools appear, so you know when live HRIS data is out of scope. It does not replace your company rubric or legal HR policy, but it standardizes narrative feedback so ratings discussions stay evidence-based. Skip it if you are a true solo operator with no reports; it shines when you must write reviews quickly without an HR department.1.6kinstalls64Journal Entry PrepJournal Entry Prep is an agent skill for solo founders and tiny teams who still own bookkeeping touchpoints at month-end even when an accountant signs off. Invoke it when you need to book accruals for received-but-unbilled services, amortize prepaids, run depreciation schedules, mirror payroll accruals, or sketch revenue-recognition entries with coherent debit and credit accounts. The skill encodes common accrual archetypes, where to pull support (open POs, contracts, recurring vendors, submitted expense reports), and what to document so a reviewer can post with confidence. It emphasizes consistent estimation, auto-reversals in the next period, and tracking actuals against accruals to improve future closes. It does not replace a CPA or controller; treat outputs as drafts for professional review. Best when you already use a formal GL chart and close calendar; skip if you are pre-revenue with cash-basis simplicity only.1.6kinstalls65Brand ReviewBrand-review is an agent skill that audits marketing and product copy against your brand voice, written style guide, messaging pillars, and terminology rules. Solo builders and small teams use it as a checker before pages, emails, ads, or docs go live, or when they need a consistency sweep across existing assets. The workflow ingests content from paste, files, linked docs, or URLs, resolves brand guidelines from local settings when present, and otherwise asks for guidelines or applies a general professional review. Each finding is classified as critical, major, or minor with specific fixes, plus separate compliance screening for unsubstantiated claims, missing disclaimers, and competitor mentions. It fits the Prism journey as a Ship-phase review gate that also supports Launch distribution copy and Grow content maintenance without replacing legal counsel for regulated industries.1.6kinstalls66Discover BrandDiscover-brand is an orchestration skill for solo builders and small teams who need a single pass over scattered brand knowledge before shipping marketing, sales, or product UI. Instead of manually opening every SaaS tool, the skill coordinates autonomous search across common enterprise connectors, triages what counts as authoritative versus stale, and produces a structured discovery report with ranked sources, noted conflicts, and questions you must answer before locking voice and visuals. It fits the moments when you are preparing a launch narrative, refreshing lifecycle email tone, or aligning a new landing page with whatever style guide already exists in Drive or Notion. The workflow deliberately orients the user first—search, analyze, then optional guideline generation—so expectations stay clear when connectors are partial or permissions are uneven. Treat outputs as an audit input: review the report, resolve conflicts, then proceed to formal brand guideline creation rather than assuming completeness from search hits alone.1.6kinstalls67Compliance TrackingCompliance Tracking is an agent skill for solo and indie builders who sell software into regulated or security-conscious buyers and cannot afford a full-time GRC team. It structures ongoing regulatory readiness instead of a panic spreadsheet the week before an audit. You work through a control inventory tied to framework requirements, an audit calendar with evidence and remediation deadlines, explicit evidence management for each control, and a gap analysis that ranks fixes on a realistic timeline. The skill references five widely used frameworks—SOC 2 for service organizations, ISO 27001 for ISMS discipline, GDPR for EU privacy, HIPAA for US healthcare PHI, and PCI DSS for cardholder data—so you can scope which regime applies without pretending one checklist covers everything. Use it when triggers like “compliance,” “audit prep,” “SOC 2,” or “regulatory requirement” appear, during security reviews before launch, and in Operate when policies or infra change and evidence must be refreshed. Outputs are organized tracking artifacts and remediation plans your agent can extend into policies, tickets, or runbooks; review any automated suggestions against your counsel and the Security Aud1.6kinstalls68Create An AssetCreate an Asset is an Anthropic knowledge-work plugin skill that guides an agent to build tailored sales and marketing collateral from structured inputs: prospect company and contacts, audience type, deal purpose (intro through close), and deliverable format. Solo founders and indie builders doing their own outbound use it when they need an impressive multi-tab landing experience, a meeting deck, a leave-behind one-pager, or a workflow demo that shows how your product connects to a prospect’s stack. Invocation is conversational via /create-an-asset or explicit prompts with transcripts and pain points. After generation, you iterate on brand colors, sections, and length without restarting the brief. It spans Validate when you need a credible landing story, Launch when distributing materials, and Grow when nurturing deals—without replacing a full design system or CRM automation.1.6kinstalls69Create Cowork PluginCreate Cowork Plugin is a meta skill for solo builders and indie developers who package procedural knowledge into installable Cowork plugins rather than one-off prompts. It walks through component schemas for skills and related plugin parts, including YAML frontmatter contracts, writing-style rules, and how to split long guidance into references. The intent is consistent, trigger-rich descriptions so Claude and similar agents invoke the right capability. Use when you need to design an API skill, document endpoint conventions, or implement any plugin component after planning in earlier phases. It does not replace product brainstorming; it standardizes how capabilities are encoded once you know what to build.1.6kinstalls70Guideline GenerationGuideline-generation is an agent skill that teaches how to assign and interpret confidence scores on brand-guideline sections produced from mixed sources—official style guides, email templates, call transcripts, and chat. Solo builders and tiny teams use it when an agent drafts a brand book and they need to know which bullets are safe to ship versus which need human sign-off. The methodology defines High confidence when multiple sources agree with authoritative backing, Medium when inference was required, and Low when evidence is thin or conflicting. It fits the Build docs phase for maintaining living brand kits, but also helps Validate when scoping positioning and Launch when aligning external copy. Pair it with document-ingestion skills first, then human review on any Low bucket before publishing to site or ads.1.6kinstalls71Recruiting PipelineRecruiting Pipeline is an agent skill that helps solo and indie builders manage an end-to-end hiring funnel without a dedicated recruiter. It maps candidates across six stages—Sourced, Screen, Interview, Debrief, Offer, and Accepted—and spells out the primary action at each step, from personalized outreach through comp negotiation and onboarding handoff. The skill emphasizes measurable recruiting health: how long candidates sit in each stage, where drop-off happens, which channels produce hires, and how often offers convert. When an applicant tracking system is connected, the agent can pull candidate records and refresh statuses automatically; otherwise it works from conversational updates and manual pipeline checks. Triggers include recruiting updates, candidate pipeline questions, hiring status, and discussions of sourcing, screening, interviews, and offers—useful when you are hiring your first engineer, designer, or support contractor and need a consistent vocabulary and metrics lens instead of scattered spreadsheets.1.6kinstalls72Synthesize ResearchSynthesize Research is an agent skill that compresses scattered user evidence into decision-ready insight. It accepts interviews, surveys, usability summaries, support tickets, sales notes, and—when connectors are available—knowledge-base documents, feedback streams, analytics funnels, and meeting transcripts. The workflow begins by cataloging what you have and what kind of research it represents, then moves toward thematic extraction and prioritization by how often pain points appear and how severely they affect outcomes. That makes it valuable for solo builders who talk to customers directly and end up with piles of notes but no single narrative for what to build next. Primary placement is Idea research, but the same synthesis loop helps during Validate when narrowing MVP scope and during Grow when support volume reveals product gaps. Connectors are optional; pasted text and uploads work standalone. Use it when raw feedback exceeds your mental spreadsheet, not when you already have a signed spec and task list.1.6kinstalls73Close ManagementClose-management is an agent skill that structures the month-end close for solo founders and tiny teams who still own the books alongside the product. It provides a pre-close window (reminders, cut-off confirmation, preliminary bank reconciliation, open PO accrual review, payroll schedule checks), then day-one activities such as sub-ledger completion, AP accruals for uninvoiced receipts, payroll posting, and cash recording. The skill emphasizes task sequencing, dependencies, and blocker identification rather than automating ledger posts—appropriate when your ERP, payroll, and banking modules must align before financials are trustworthy. Invoke it when planning the close calendar, tracking in-flight close tasks, or reordering work when AP or payroll slips. All outputs are assistive; the readme states clearly that qualified professionals must review close entries and judgments.1.6kinstalls74Contact ResearchContact-research is an interpretation guide for contact and account signals that solo builders and tiny GTM teams use when deciding who to email, who needs support empathy, and who might be a champion. It organizes activity signals (opens, replies, meetings, community posts, tickets, doc visits, logins, events), relationship health (dormancy windows and cross-channel engagement), and professional moves (job changes, promotions, decision-maker titles) into plain-language guidance rather than raw CRM fields. Install it when your agent is summarizing a person or account before you draft outreach, plan a QBR, or choose a re-engagement angle. It does not pull live CRM data by itself; it teaches how to read signals once your tools have surfaced them. Pair with your existing stack (email, product analytics, support) so recommendations stay grounded in evidence.1.6kinstalls75Roadmap UpdateRoadmap Update is an Anthropic knowledge-work plugin skill that helps solo and indie builders add initiatives, change statuses, reshuffle priority, and move dates without losing the story of why. It starts by syncing or importing your current plan—via a connected project tracker when available—and surfaces overdue, blocked, and ownerless items before you edit anything. You pick an operation (new item, status change, reprioritize, or timeline shift), and the skill gathers the missing fields, suggests placement against capacity, and runs an explicit tradeoff discussion so something has to give when you add scope. The workflow ends with a concise change summary and updated Now/Next/Later or equivalent view you can drop back into your PM tool or shared doc. Use it when priorities flip after validation, a dependency slips mid-build, or you are standing up a roadmap from scratch and want agent-guided structure instead of ad-hoc list edits in chat.1.6kinstalls76Compose OutreachCompose Outreach is an agent skill that turns rough prospect context into ready-to-send sales touchpoints using Anthropic knowledge-work plugin guidelines. Solo and indie builders use it when they need founder-led outbound without sounding generic: short emails that open on a specific signal, call guides instead of robotic scripts, and LinkedIn intros that respect character limits. The skill encodes format rules—subject length, sentence caps, and a single ask—so your agent does not ramble or bury the CTA. It fits SaaS and content businesses where you still close deals yourself and want repeatable messaging templates the agent can fill from CRM notes or research. Pair it with your own hook research; it does not replace list building or compliance review of CAN-SPAM and LinkedIn policies.1.6kinstalls77Cowork Plugin CustomizerCowork Plugin Customizer is an agent skill backed by a manifest fragment that lists production-style MCP server URLs and auth patterns plus a curated set of plugin categories for knowledge-work setups. Solo builders shipping with Claude Code or similar agents use it when they want a starting map of which remote MCP endpoints to enable—version control, project management, chat, design, and observability—without hand-typing every transport type and header name. The skill does not replace secret management: you still supply tokens via environment variables. It accelerates the build-phase integration pass where your agent needs Figma, Slack, or Datadog in one coherent plugin profile. Treat it as a scaffold to edit for your stack, then validate each server in a sandbox before granting broad tool access in production repos.1.6kinstalls78Enrich LeadEnrich Lead is an Apollo-backed agent skill that converts whatever you know about a prospect—first name, domain, LinkedIn profile, or work email—into a structured contact dossier your agent can act on. Solo and indie builders shipping B2B or prosumer products install it when cold outreach, partnership pings, or investor intros need verified titles, company intel, and reachable channels instead of guesswork from a single Google tab. The workflow parses $ARGUMENTS for every identifier, resolves ambiguous inputs (for example CEO of Figma) with a filtered people search, then runs people match with optional personal email reveal while warning that enrichment spends one Apollo credit. It fits Claude Code setups that already expose the Apollo MCP tools named in the skill body. Use it in the Grow phase when lifecycle and sales motions need fast enrichment; skip it if you are not using Apollo or cannot accept per-match credits.1.6kinstalls79Org PlanningOrg Planning is an agent skill for headcount planning, organizational design, and team-structure optimization when you are thinking about team size, reporting lines, or who to hire next. It walks four planning dimensions—how many people in which roles, reporting structure and spans, critical hire order, and headcount cost trade-offs—and applies published healthy-range benchmarks so you can spot top-heavy orgs or broken span of control before they slow decisions. Solo and indie builders use it when moving from founder-only delivery to first employees, planning a light reorg, or pressure-testing whether a proposed structure matches the work you still need to ship. The skill produces text-based org charts, sequenced hiring roadmaps, and cost-oriented headcount plans, and explicitly calls out structural risks like single points of failure. It is methodology and modeling in chat, not HRIS or payroll integration.1.6kinstalls80Sprint PlanningSprint Planning is an agent skill to scope work, estimate team capacity, set sprint goals, and draft a sprint plan when you are kicking off a new iteration. The standalone path always works: define success criteria, subtract PTO and meeting load from capacity, prioritize backlog items, surface dependencies and risks, and generate a sprint plan document. A supercharged mode can pull backlog and create sprints in connected project trackers and read calendar availability when CONNECTORS are wired. Solo builders and tiny teams use it to avoid over-committing a one-person sprint, to separate P0 from stretch when launch pressure hits, and to account for carryover from the last sprint without rebuilding the plan from scratch. Invoke with /sprint-planning and an optional sprint name or date range. It complements org-level headcount planning by operating at the iteration level.1.5kinstalls81Prospectprospect is a user-invocable agent workflow that accepts an ideal customer profile in everyday language—like VP Engineering at Series B US SaaS—and returns structured, enriched leads ready for outreach. Solo founders who hate bouncing between search UIs install it when they need volume and contact data without hiring an SDR. The skill decomposes the prompt into Apollo-oriented filters, validates ambiguity with at most two questions, searches organizations, then ranks people with emails and phone numbers. It assumes integrated sales-intelligence tools in the agent environment rather than manual CSV juggling. Best paired with clear title and segment constraints; vague “any startup” asks get narrowed before spendy API calls.1.5kinstalls82Write QueryWrite-query is a knowledge-work agent skill that translates a natural-language description of the data you need into production-oriented SQL tuned for your engine. The workflow first decomposes the ask—columns, time windows, segments, aggregations, join paths, sort order, and row limits—then selects or confirms the dialect (PostgreSQL family, Snowflake, BigQuery, Redshift, Databricks SQL, MySQL, and similar). It favors readable multi-CTE layouts and dialect-specific best practices, including attention to heavy partitioned tables when optimization matters. Solo builders shipping SaaS or content products use it to answer growth questions and operational checks without memorizing vendor quirks, while still getting queries they can paste into their BI tool or run against Supabase, Neon, Aurora, or cloud warehouses.1.5kinstalls83Performance ReportPerformance Report is an agent skill for solo and indie marketers who need more than a spreadsheet dump when closing a campaign or preparing weekly, monthly, or quarterly updates. It walks you through choosing report type—single campaign, one channel, content portfolio, or cross-channel overview—then aligns the reporting window and data sources before drafting narrative insights stakeholders can act on. When marketing analytics connectors are wired in, the skill can discover accounts and ingest performance data automatically; otherwise you can supply exports or manual figures and still get the same report skeleton. The emphasis is translation: raw KPIs become trend context, explicit wins and misses, and a prioritized optimization backlog for the next period. It fits builders shipping content products, SaaS with paid or organic motion, or small teams without a dedicated analytics writer—anyone who must justify spend and plan the next sprint from evidence rather than vibes.1.5kinstalls84Status ReportStatus Report is an agent skill from Anthropics knowledge-work plugins that drafts polished leadership updates from whatever project context you provide—tracker exports, bullet notes, or chat summaries. Solo and indie builders use it when they need a repeatable narrative for advisors, cofounders, or clients without spending an afternoon reformatting spreadsheets. The workflow centers on /status-report with optional period and team arguments and produces a consistent markdown brief: executive summary, rollup health, metric scorecard, wins, active work with owners, and a risks table aligned to stakeholder decision-making. It explicitly points to risk-assessment for deeper risk-matrix language when impact or severity need formal definitions. The skill fits anyone wearing the PM hat alongside engineering: you supply raw facts; the agent structures them for scanability and escalation. Pair it with your real numbers from analytics or issue trackers; it does not invent installs or audit results.1.5kinstalls85Vendor CheckVendor-check is an Anthropic knowledge-work plugin command that aggregates a single vendor’s contractual posture across every connector you have wired up—contract lifecycle management, CRM, email, and document repositories. Solo founders and small teams use it when onboarding a new supplier or preparing a renewal, when they need to know which core documents actually exist versus which are still missing, and when expiration dates or post-termination obligations might slip through siloed tools. The workflow starts by normalizing the vendor identity, then queries connected systems and synthesizes status, gaps, and upcoming deadlines into a report meant for human verification. It explicitly does not replace counsel; it reduces thrash from manually opening four apps and guessing whether a DPA was ever executed.1.5kinstalls86RunbookRunbook is an operational documentation skill that produces repeatable, on-call-ready procedures from a task name or process description. Solo builders and tiny teams use it when a workflow still lives in chat history or one senior engineer’s head—deploy refreshes, credential rotations, data fixes, incident mitigations—and needs exact commands, expected outcomes, and what to do when a step fails. Output follows a fixed markdown skeleton with prerequisites, step blocks, verification, troubleshooting matrix, and ownership metadata so the next person can execute without a walkthrough. It fits after you ship something that must be operated, not while you are still sketching features. CONNECTORS.md is referenced when placeholders or tool integrations are unclear, keeping runbooks honest about external systems.1.5kinstalls87Process Docprocess-doc is a knowledge-work agent skill that documents business processes as standard operating procedures a solo builder can hand to a contractor, co-founder, or future self. Invoke it when a workflow lives only in someone's head, when you need a RACI to end "who owns this?" confusion, or when audits and handoffs require written steps with triggers, outputs, and exceptions. The skill walks you through the process via conversation—describe it, paste fragments, or supply a name and answer targeted questions—then emits a structured markdown SOP with purpose, scope, RACI table, flow visualization, and numbered steps with roles and timing. It fits across the journey: scoping how validation work runs, documenting launch checklists, support escalations, or billing operations. Output is planning-grade prose suitable for Notion or the repo docs folder, not executable automation. Pair it with connector docs when placeholders reference external tools.1.5kinstalls88Comp AnalysisComp-analysis is an agent skill for analyzing compensation—benchmarking roles, placing offers in bands, and modeling equity grants. It is built for solo founders and small teams who need defensible pay answers without a full People Ops stack. Invoke it when you ask what to pay a role, whether an offer is competitive, or when you upload comp data to spot outliers and retention risk. The skill structures total compensation into base, equity, bonus, and benefits and factors in level, geography, and company stage. Outputs support hiring negotiations, refresh planning, and band hygiene. It assumes you can supply a role description, location, or a CSV of internal bands; connector placeholders refer to optional external data sources documented separately.1.5kinstalls89Draft Contentdraft-content is an Anthropic knowledge-work agent skill that produces marketing drafts tailored to content type, audience, and brand voice. Solo and indie builders use it when they need blog posts, platform-specific social copy, newsletters, landing-page text, press releases, or case studies without starting from a blank page. The workflow asks for content type, topic, target reader, key takeaways, tone, and length when those inputs are missing, then generates structured copy with headline or subject-line options where appropriate. It fits the Grow and Launch phases of the Prism journey: launch teams adapt messages for SEO and distribution surfaces, while growing products rely on consistent content cadence. The skill is procedural prose in SKILL.md form—invoked inside Claude Code, Cursor, or similar agents—not a hosted CMS or MCP server. Connectors referenced in CONNECTORS.md may extend reach but the core value is disciplined, repeatable drafting aligned to marketing channels.1.5kinstalls90ForecastForecast is an agent skill for solo and indie builders who still own sales numbers and need a defensible quarterly view without a full RevOps stack. You supply a period, optional pipeline file, and quota context; the skill walks through weighted projections by stage, separates commit from upside, and surfaces gap analysis when coverage is thin. It is aimed at founders preparing forecast calls, deciding which deals to commit, or sanity-checking a CSV export before updating investors or the team. CONNECTORS.md applies when placeholders or live CRM tools are in play, but the documented path works from a standalone upload. Compared with ad-hoc spreadsheet tweaks, it structures scenario thinking and narrative-ready breakdowns agents can reuse each quarter.1.5kinstalls91Call SummaryCall Summary is an agent skill for solo founders and small sales-led teams who finish calls with messy notes and need consistent outputs fast. You pass call notes or a transcript—optionally via file reference—and the workflow extracts decisions, action items with owners and dates, objections, and open questions. It then drafts a customer follow-up email and an internal summary suitable for CRM logging or team channels. CONNECTORS.md covers optional tool hooks, but the documented standalone path always works from pasted content. Use it to reduce the gap between hanging up and sending a professional recap, especially on discovery, demo, or negotiation calls where dropping follow-up means lost deals.1.5kinstalls92Weekly Prep BriefWeekly Prep Brief structures how an agent researches and formats preparation for a week of customer conversations. It scales effort to account type—strategic logos, open opportunities, renewals inside ninety days, mid-market routines, and lightweight check-ins—so a solo founder wearing sales and success hats does not over-research low-stakes calls or under-prepare for revenue risk. The skill prioritizes week-level signals such as trial expiry, renewal windows, critical tickets, expansion cues, executive attendance, and competitive or champion job changes. When internal systems like Common Room are empty, it tells the agent to say so, lean on external research, and suggest offline team context. For heavy weeks it offers prioritization when seven or more meetings stack up. The output is meeting-ready narrative context, not code—aimed at indie SaaS operators running their own pipeline.1.5kinstalls93Metrics ReviewMetrics Review is an agent skill for solo founders who have traction data but need a repeatable ritual instead of staring at dashboards. It walks through gathering key product metrics for a chosen period, lining up comparisons (prior period, year-ago, targets), and capturing context like launches or incidents. When a product analytics connector is available it pulls segments automatically; otherwise you paste tables or describe the numbers. The skill prompts for review scope—full suite versus a focused metric—and then organizes the data into trends and recommended actions suitable for a weekly, monthly, or quarterly review. Use it in Claude Code or similar agents when a metric moved sharply or you owe yourself a honest scorecard before the next build cycle.1.5kinstalls94Vendor ReviewVendor Review is a knowledge-work agent skill that walks a solo builder through a disciplined vendor decision using the `/vendor-review` command. Whether you are comparing two observability tools, deciding whether to renew an auth provider, or sanity-checking a proposal before you charge it to the company card, the skill forces total cost of ownership, operational risk, and performance evidence into one narrative. It asks for vendor name, context, and contract or proposal details, then applies a consistent framework: hidden migration and exit costs, SLA and support behavior, security posture, and dependency concentration. The output is recommendation-ready analysis with negotiation hooks—not a generic pros-and-cons list. It pairs well with Prism’s validate journey when pricing and scope still flex, and with operate when renewals arrive. Connectors referenced in CONNECTORS.md may enrich answers when your workspace links billing or CRM tools.1.5kinstalls95Email SequenceEmail Sequence is a knowledge-work agent skill invoked with `/email-sequence` to produce complete drip campaigns instead of single-message drafts. Solo builders launching SaaS products routinely need onboarding that activates, nurture that converts trials, and win-back paths that respect unsubscribe psychology—the skill collects sequence type, goal, audience stage, and optional length, then outputs timed messages with branching and exit rules. Sequence types span onboarding, lead nurture, re-engagement, product launch, event follow-up, upgrade paths, win-back, and educational drips. Each run is meant to be handed to Mailchimp, Customer.io, or similar with benchmarks and A/B ideas so you can measure opens and conversions against stated goals. It sits firmly in the Grow journey because it compounds users after Ship and Launch, though launch announcement arcs can borrow the same framework. See CONNECTORS.md if your agent workspace links email or CRM connectors for richer personalization context.1.5kinstalls96Stakeholder UpdateStakeholder Update is an agent skill for solo and indie builders who need consistent status communication without maintaining five different templates. You choose among weekly, monthly, launch, or ad-hoc updates, then specify whether the reader is leadership, engineering, cross-functional partners, customers, or the board; the skill produces framing, metrics emphasis, and jargon level matched to that pairing. It fits the build phase when you are running a product as your own PM, but the same skill supports launch announcements, growth-facing timelines, and operate-phase escalations when blockers or risks need a formal narrative. Compared with ad-hoc chat, it enforces audience-aware structure so updates stay outcome-focused for execs, implementation-clear for engineers, and benefit-led for external readers. Connectors referenced in the plugin docs may supply live project context; the core value is repeatable stakeholder messaging across cadences and audiences from one invocation.1.5kinstalls97Kb Articlekb-article is an Anthropics knowledge-work plugin skill that drafts publish-ready knowledge base articles from material you already have—a closed ticket, a workaround, or a question that keeps hitting support. It is built for solo builders and small teams running SaaS who need consistent KB structure without hiring a technical writer for every resolution. The workflow starts by extracting the original problem, the fix, who is affected, and whether the issue is recurring, then picks an article type such as how-to, troubleshooting, FAQ, known issue, or reference. Examples in the skill cover SSO setup, export limits, webhooks, and browser-specific bugs. Use it after you have a verified resolution; pair with your connector docs when the draft must reference live systems. The output targets customer self-service and search, not agent coding plans.1.5kinstalls98Briefbrief is an Anthropic knowledge-work legal briefing skill with three invocation modes—daily summary, targeted topic research, and incident response.context assembly. Small teams with legal exposure use it when starting the day, researching an internal legal question, or reacting to a developing situation such as a data breach, litigation threat, or regulatory inquiry. The workflow scans connected sources like email and calendar for contract and compliance signals, synthesizes a structured briefing, and insists that qualified legal professionals review output before reliance. For Prism's solo-builder audience, it is most relevant when you wear general counsel-lite responsibilities at a startup with connectors enabled; it is not a substitute for counsel. The skill is procedural agent guidance tied to CONNECTORS.md integrations, spanning Operate monitoring for daily and incident use and Validate-style scoping when researching a named legal topic.1.4kinstalls99OnboardingOnboarding is an agent skill that generates a comprehensive new-hire checklist and first-week plan for indie founders and small teams who finally have a start date to plan around. Invoke it when someone is joining soon, you need a pre-Day-1 task list for accounts and equipment, you want Day 1 and Week 1 scheduled with a buddy, or you are setting 30/60/90-day goals for a new contributor. The skill walks through required inputs—name, role, team, start date, and manager—and returns structured markdown with pre-start todos, a timed Day 1 table, and escalation paths for IT and manager touchpoints. It fits Prism’s operate phase because hiring is an operational inflection for solos shipping a product, not a coding integration. Pair it with your HR or ops docs and connected workplace tools after checking CONNECTORS.md.1.4kinstalls100Sequence LoadSequence Load is a user-invocable Apollo workflow skill for solo founders and small GTM teams who run outbound from Claude with the Apollo MCP connected. It parses "$ARGUMENTS" into Apollo search filters such as person_titles, person_seniorities, q_organization_keyword_tags, employee ranges, and locations, then resolves the target emailer campaign by name, handles enrichment and contact creation with deduplication, and bulk-adds prospects to the sequence. Examples cover listing campaigns, loading 20 VPs into a Q1 sequence, and reloading more leads into an enterprise pipeline. Install it when you want one slash-style command instead of manually clicking through search, enrich, and enrollment in Apollo.1.4kinstalls101Digestdigest is a journey-wide agent skill from the knowledge-work plugins that produces a daily or weekly activity digest across every MCP source you connect—chat channels and DMs, email threads, cloud docs, project trackers, CRM updates, and wiki pages. Solo builders use it when catching up after time away, starting the day with a summary of mentions and action items, or reviewing a week’s decisions and document changes grouped by project. The agent parses --daily, --weekly, or --since ranges, checks which connectors are live, and emits a structured highlight of what matters instead of manually opening six apps. It depends on MCP configuration documented alongside CONNECTORS.md and shares the same source-detection approach as the search command. Because the ritual is timing-driven rather than phase-specific, it remains useful across build PM, operate iterate, and grow support contexts whenever connected tools hold your workflow state.1.4kinstalls102Capacity PlanCapacity-plan is an Anthropics knowledge-work agent skill for resource and workload planning. Solo builders and tiny teams use it when quarterly planning looms, everyone feels overallocated, or you need numbers to decide whether to hire, cut scope, or push timelines. The workflow asks for team roles, current workload (from a tracker or narrative), upcoming work, and constraints such as budget and hiring lead time. It structures analysis across people, budget, and time—including utilization targets by role type—so allocation decisions are explicit rather than gut feel. It fits indie operators who wear PM and engineering hats and need a repeatable capacity narrative before locking roadmaps. Invoke it early in planning cycles; pair outputs with prioritization and hiring conversations rather than treating utilization percentages as law.1.4kinstalls103Change Requestchange-request is an agent skill that turns a vague “we should change X” idea into an approval-ready change record for solo builders and small teams who still need disciplined rollouts. It walks you through what is changing, who is affected, expected resistance, and how significant the change is, then forces communication, training, and support plans—not just a technical diff. The execute and sustain sections keep adoption and reinforcement on the page so deployments do not end at merge. Use it when proposing a system or process change that needs sign-off, preparing material for change-advisory review, documenting risk and rollback before deployment, or planning stakeholder messaging for a rollout. It pairs well with infra and release workflows where explaining why matters as much as what ships.1.4kinstalls104Pipeline Reviewpipeline-review is an agent skill for solo founders and small sales teams who need a structured weekly pipeline review without standing up a separate analytics product. Invoke it with an optional segment or rep hint; it accepts CSV exports or pasted deal tables and walks through health checks, risk flags, prioritization, and hygiene issues such as unrealistic close dates or single-threaded accounts. The skill is phase-specific to Grow because its output is operational focus for the next week—not product scope or code changes. It pairs with knowledge-work connector docs when CRM integrations exist but deliberately works standalone. For indie SaaS builders wearing sales hats, it turns raw CRM dumps into ranked actions instead of ad-hoc spreadsheet sorting. Use it when deciding which deals deserve outreach this week or when auditing pipeline quality before a forecast conversation.1.4kinstalls105Journal EntryJournal Entry Preparation is an agent skill for solo builders and small teams who own their own books and need consistent, reviewable GL drafts at period close. It walks through gathering source data (optionally from ERP or warehouse connectors), building properly balanced entries for AP accruals, fixed-asset depreciation, prepaid amortization, payroll accruals, and revenue or deferred-revenue adjustments, and packaging supporting detail auditors expect. Use it when month-end is approaching and ad-hoc chat produces unbalanced or undocumented postings. The skill emphasizes professional review before anything is posted and fits operators who already have chart-of-accounts discipline but want procedural help so accruals do not get skipped under shipping pressure.1.4kinstalls106Draft ResponseDraft Response is a knowledge-work plugin skill that walks an agent through writing professional, customer-facing messages tailored to urgency, channel, and relationship. Solo SaaS founders use it when replying to product questions, multi-day integration outages, feature declines, or billing errors—cases where a rushed Slack paste can damage trust. The skill defines a /draft-response workflow that first classifies the customer, situation type, urgency, and channel before composing copy at the right formality. It expects optional connector context from the parent plugin ecosystem for account lookup, but the core value is consistent voice and empathy under pressure. It spans Grow support as the primary shelf but also applies when Ship or Launch incidents require outbound status communication. Not a CRM replacement; it produces draft text you review before send.1.4kinstalls107People ReportPeople Report is a knowledge-work plugin skill that generates structured people-analytics reports from HR systems. Solo founders who have grown into a small team—or indie CEOs wearing the people-ops hat—use it for leadership snapshots: current headcount by team and level, voluntary versus involuntary turnover trends, diversity representation, and org-health signals like management layers and flight risk. The /people-report command accepts a report type argument and maps to defined metric families spanning retention, diversity, engagement, and productivity. It is Operate-phase tooling for iterating on hiring plans, reorg sizing, and risk conversations, not for validating a product idea with no employees. Effectiveness assumes connected HR data sources documented in the plugin connectors guide.1.4kinstalls108Customer EscalationCustomer-escalation is an agent skill from Anthropic’s knowledge-work plugins that packages live support problems into engineering- and leadership-ready briefs. Solo builders and tiny teams wearing support hats use it when a bug outgrows tier-one troubleshooting, the same defect lands from multiple accounts, a customer is near churn, or an ticket has blown past its SLA. The workflow starts by clarifying the core failure mode, affected customers or segments, how long the issue has persisted, and what was already attempted. It then shapes reproduction steps, business impact, and the right escalation target so engineering or product can prioritize without re-interviewing the customer. Connectors referenced in CONNECTORS.md may extend context gathering, but the core value is consistent escalation narrative—not ad-hoc Slack threads.1.4kinstalls109Draft OfferDraft-offer is a knowledge-work plugin skill that helps solo founders and small-team hiring managers produce polished offer letter drafts without opening a separate HR template stack. Invoke it when a candidate is at the offer stage and you need to assemble base salary, equity with vesting, optional signing bonus, target bonus, and first-year total compensation in one readable package. The skill walks through missing inputs—title, level, location, start date, and reporting manager—and emits structured markdown with a compensation table, employment terms, and a benefits summary block you can paste into DocuSign or counsel review. It can also frame negotiation guidance for the hiring manager when comp bands are still fuzzy. It does not replace legal review or jurisdiction-specific employment law; it accelerates consistent drafting so you spend less time reformatting numbers in email.1.4kinstalls110Triage Ndatriage-nda is a knowledge-work plugin skill that rapid-screens non-disclosure agreements so operators can decide whether an NDA can move on standard approval, needs counsel review, or requires full legal review. It is built for the moment a new NDA arrives from sales or business development—not for drafting contracts from scratch. The workflow accepts uploads, links, or pasted text, then loads organization-specific criteria from local settings such as a legal.local.md playbook covering mutual obligations, term length, carveouts, and prohibited clauses. Classification maps to GREEN for routine signing within delegation, YELLOW when a lawyer should sanity-check edge terms, and RED when negotiations or full review are mandatory, including attention to sneaky non-solicit or non-compete language. Solo founders doing their first enterprise pilots may still use it as a structured checklist, but the skill assumes you maintain screening rules and will have qualified professionals validate outcomes. It assists legal operations workflows without offering legal advice, aligning with Prism’s audience when small teams suddenly face procurement NDAs alongside product work.1.4kinstalls111Signature RequestSignature-request is an agent skill that prepares finalized agreements for electronic signature. Solo and indie builders use it when an MSA, order form, or contractor agreement is ready to execute—not while still redlining. The workflow accepts the document from a file, link, or reference, walks through a structured pre-signature checklist (no open edits, attachments, correct signatories, execution dates), then sets signing order and routes for execution. It fits operators and founders who close deals without a full legal ops team but still need fewer envelope mistakes. Pair it with your CLM or e-sign connector documentation; the skill does not substitute for counsel on whether the contract terms are right, only that the package is complete before send.1.4kinstalls112Sox TestingSox-testing is an agent skill from Anthropic’s knowledge-work plugins that structures Sarbanes-Oxley Section 404 testing chores. Finance and ops leads use it when planning quarterly or annual tests: pulling samples for revenue, procure-to-pay, payroll, close, treasury, fixed assets, inventory, or IT general controls, then scaffolding workpapers and deficiency narratives. It is built as a slash-command workflow with argument hints, not as a replacement for auditors or counsel. Solo builders rarely need it unless they wear a controller hat at a SOX-scoped company; small finance teams install it to speed consistent templates while keeping human sign-off on every workpaper destined for audit files. The skill emphasizes document generation and assessment framing tied to named control areas rather than generic checklist chat.1.4kinstalls113Scientific Problem SelectionScientific Problem Selection is an agent skill from Anthropics knowledge-work-plugins line, intended to help you pick which scientific or technical problem deserves your next cycle of attention. Solo builders and small teams doing ML, bio, or hard-tech adjacent work often fail not from poor execution but from committing to a question that is ill-posed, already solved, or intractable with their constraints. This skill sits at the very start of the journey: after you have a fog of interesting angles but before you run costly experiments, procure data, or wire an agent pipeline around the wrong objective. Prism’s ingested excerpt is license text only, so treat capability bullets as inferred from the skill name and collection context until the full SKILL.md is available. Expect structured prompts or criteria for impact, feasibility, and novelty rather than boilerplate code generation.1.2kinstalls114Single Cell Rna QcSingle-cell RNA QC is an agent skill in Anthropics knowledge-work-plugins line aimed at researchers and solo technical founders running single-cell RNA sequencing pipelines. Prism ingestion surfaced only the Apache 2.0 license text in the readme snippet, so operational steps should be confirmed in the full SKILL.md inside the repo before relying on thresholds or tool commands. By name and catalog placement, the skill supports quality control—filtering low-quality cells, spotting batch effects, and documenting pass/fail criteria—before you ship notebook outputs or API-backed analysis to stakeholders. Tag it phase-specific under Ship/testing for data-heavy products where bad QC silently poisons launch claims. Confidence is moderate until the procedural body is visible in catalog sync.1.1kinstalls115Nextflow Developmentnextflow-development is an agent skill aimed at builders and researchers who maintain Nextflow pipelines—declarative workflow files that chain processes over channels for genomics, proteomics, and other data-heavy batch jobs. Prism lists it from the anthropics knowledge-work-plugins bundle; the indexed readme in ingest is license text only, so treat operational detail as coming from the full SKILL.md in your checkout. In typical use, the skill should guide module layout, process boundaries, resume/cache strategy, executor configuration, and testing patterns so pipelines stay reproducible across laptops, Slurm, and cloud batch. Solo devs in biotech startups or indie science tooling use it when extending nf-core-style workflows or authoring internal DSL2. It is not a general CI skill or a substitute for domain-specific biology QC rules—pair it with your org’s validation and compliance requirements.1.1kinstalls116Instrument Data To AllotropeInstrument Data to Allotrope is an agent skill from Anthropics’ knowledge-work plugins that helps solo builders and small lab-adjacent teams turn instrument exports into Allotrope-aligned structures. Allotrope Foundation formats matter when you need interoperable analytical chemistry and measurement metadata instead of one-off CSV quirks. Use this skill when you are building ingestion for R&D tools, LIMS bridges, or internal dashboards that must speak a shared scientific data model. The skill positions your coding agent to apply domain conversion steps during integration work—reducing manual reformatting and documentation drift. It is narrower than a general ETL framework: it targets Allotrope-oriented instrument pipelines. Pair it with your own validation rules and storage (object store, database, or API) once the canonical shape is defined. Expect intermediate familiarity with lab data concepts and file formats; the agent accelerates mapping logic and repeatable transforms rather than replacing domain sign-off.1.1kinstalls117Scvi Toolsscvi-tools is an agent skill oriented around the scVI ecosystem for single-cell and spatial omics: variational autoencoder models, integration across batches, and downstream statistics on count matrices stored in AnnData. Solo and indie builders who ship bioinformatics features, internal research tools, or data-heavy ML products install it when they want the coding agent to follow library-native APIs rather than inventing ad-hoc PCA pipelines. Use it during Build when you are implementing notebooks, ETL into feature stores, or reproducible training jobs that must cite scVI, scANVI, totalVI, or related model constructors correctly. It matters because single-cell work fails silently when normalization, layer keys, and train/validation splits are wrong—the skill nudges agents toward scvi-tools conventions. Published packaging in this repo snapshot is thin (license text only), so pair it with your own dataset contracts and pinned dependency versions before production.1.1kinstalls118Zoom McpZoom Mcp is an agent skill that explains how solo builders connect coding agents to Zoom through Model Context Protocol—using Zoom’s hosted Streamable HTTP or SSE endpoints rather than bespoke scrapers. It fits when you want agents to search meetings semantically, pull recordings and related assets, or create Zoom Docs from Markdown, while treating the live `tools/list` response as the source of truth for tool names, schemas, and newly shipped capabilities. The skill deliberately contrasts MCP’s semantic retrieval shape with deterministic meeting CRUD, steering create/update/delete flows to Zoom’s REST API when MCP does not expose them. Authentication is framed around user OAuth for the plugin’s bundled servers, which matters for indie teams who cannot rely on opaque service accounts. Pair it with your MCP-capable client (Claude Code, Cursor, or similar), validate OAuth scopes, and keep Whiteboard workflows on the dedicated whiteboard skill when boards—not meetings—are the artifact.893installs119Design Mcp WorkflowDesign MCP Workflow is a planning skill for solo builders adding Zoom (or analogous SaaS) to MCP-capable clients such as Claude. It walks through whether a task needs agentic tool use, deterministic REST automation, or both, then routes follow-up work to companion skills for zoom-mcp, rest-api, and optional whiteboard MCP paths. The workflow stresses connector expectations, hybrid architectures, and explicit assumptions about transport and authentication rather than treating MCP as a drop-in for every API call. It is not user-invocable by default in the plugin metadata, so agents typically load it when the user is designing tool-based AI workflows or asking where MCP fits. Use it during integration design on the build phase before writing handlers or registering tools, especially when whiteboard or meeting flows blur API and conversational boundaries.886installs120Contract ReviewContract Review is a lightweight legal-assist workflow for SMB founders and solo operators who lack in-house counsel but still face NDAs, MSAs, and vendor paper. The skill activates when you ask to review an agreement, surface red flags, or sanity-check payment and liability terms. It pulls documents from Gmail, DocuSign, or local files, insists on reading the complete text (including exhibits), then walks eight risk categories to produce a severity-ranked summary, plain-English explanations, and actionable negotiation guidance. A marked-up DOCX redline gives you something concrete to send back to counterparties. This is decision support—not a substitute for licensed legal advice—so it fits the Validate phase when you are deciding whether to sign, renew, or renegotiate. It can reappear in Operate when recurring vendor contracts come up for renewal. The skill pairs document ingestion references with structured analysis so agents do not skim dangerous clauses buried in schedules.869installs121Build Zoom Rest Api AppBuild Zoom REST API App is an integration skill for indie builders and small teams wiring Zoom into schedulers, bots, or internal tools. It focuses on API architecture: versioned REST under /v2, the separate GraphQL endpoint, and how OAuth token responses expose api_url so clients hit the correct regional host for data residency. The readme walks through constructing regional base URLs, vanity account patterns, and request conventions rather than a single copy-paste SDK tutorial. Use it in Build when you are authenticating with Zoom, listing meetings, or calling GraphQL and need the agent to respect EU, UK, Singapore, and other region tables instead of hard-coding the global URL. It pairs with your own OAuth app credentials and error handling; it does not replace Zoom’s official SDKs or admin consent flows. Treat it as a correctness layer so agents do not ship clients that violate residency or mix API versions.853installs122Plan Zoom ProductPlan Zoom Product is an agent skill from the Anthropic knowledge-work-plugins set for solo builders who have a Zoom-shaped idea but not yet a technical bet. The /plan-zoom-product workflow forces clarity: name the real user goal, classify whether you need automation, embedded meetings, custom video pipelines, in-meeting app behavior, reliable event delivery, AI agent tooling via MCP, or telephony and contact-center flows. When the request is ambiguous, the skill allows one short clarifier before locking a recommendation—enough discipline for a one-person team without enterprise committees. Deliverables read like a mini architecture review: primary Zoom surface, minimum supporting pieces, why weaker alternatives fail this case, ordered implementation sequence, and pointers to follow-on skills such as choose-zoom-approach. It is not a credential or OAuth tutorial; it prevents building on Meeting SDK when Webhooks plus REST would ship faster, or vice versa. Use early in validation and again when pivoting integration strategy mid-build.851installs123Setup Zoom Oauthsetup-zoom-oauth is an agent skill for solo builders who need Zoom meetings or webinars inside a SaaS or internal tool. It treats authentication as the gating design decision: who authorizes whom, which OAuth grant fits that model, and which scopes are strictly required for the user journey you are shipping. The skill walks through a short ordered workflow so you do not jump into REST endpoints while still guessing app credentials or refresh behavior. It emphasizes minimum scopes, token lifecycle including refresh rotation, and separating configuration mistakes from downstream API errors. Use it when Zoom is the blocker at credential setup, when you are choosing between grant types, or when token refresh bugs look like mysterious 401s. It is a procedural companion to Anthropic knowledge-work Zoom docs, not a drop-in OAuth library.849installs124Scribescribe is an agent skill that documents how to authenticate and call Zoom’s Build-platform Scribe transcription API when you wire meeting or media audio into your app or internal tooling. Solo builders adding voice notes, interview capture, or support-call logging need a precise map of JWT issuance, request shapes, and when to pick fast synchronous transcription versus asynchronous batch jobs. The skill explains generating HS256 tokens with the issuer set to your API key, short expiry windows, and the reality that Zoom’s docs may label the same credential pair differently across pages—so you verify labels in the live portal before shipping. Fast mode suits single short files and interactive UX via POST /transcribe with file plus config payload. Batch mode fits longer media libraries through POST /jobs and polling or webhooks. Configuration flags such as language, diarization, profanity filtering, timestamps, and channel separation are enumerated so agents do not invent unsupported fields. Use it while implementing the integration layer, not as a substitute for Zoom’s current OpenAPI reference or your own retention and consent policies for recorded audio.847installs125Probe Sdkprobe-sdk packages a disciplined preflight for Zoom Probe SDK web diagnostics so solo builders do not skip lifecycle and permission steps when audio, video, or network checks fail mysteriously. It separates meeting-runtime concerns from standalone device and network diagnostics, lists required device IDs, and enforces order from requestMediaDevicePermission through startToDiagnose and cleanup. Renderer strategy (video-tag vs canvas) and optional JS/WASM URL overrides are decision checkpoints early in the run. The skill warns that API names drift by version and points maintainers to validate against current SDK reference. Operators use it right before ship or when upgrading SDK versions to avoid leaked streams, blocked UI threads on stats listeners, and missing report snapshots.845installs126Build Zoom Meeting Sdk AppBuild Zoom Meeting SDK App is a knowledge-work plugin skill that compresses Zoom’s Android Meeting SDK integration into an implementable sequence for solo mobile builders. It assumes you are embedding live meetings inside your app—not just scheduling links—and walks through initialize and authenticate, achieve a first successful join with the default UI, then layer settings, error handling, and observability before opting into custom UI where product differentiation demands it. The guide documents a clean architecture split between Activities or Compose wrappers, an app-side meeting state machine, the mobilertc.aar facade, and a backend that issues short-lived signatures or JWTs so secrets never ship in the client. Validated doc paths and a pinned local SDK package snapshot reduce doc roulette when your agent generates integration steps. Complexity is intermediate: Android familiarity is required, and Zoom’s callback-heavy SDK rewards disciplined logging from day one.841installs127Rivet SdkThe rivet-sdk skill documents how Zoom's Rivet layer bundles auth, webhooks, and typed API clients so solo builders shipping Zoom Marketplace apps do not reinvent token refresh, signature verification, and per-product REST paths. It is aimed at indie SaaS authors adding Team Chat bots, meeting automations, or Video SDK adjacency in Node or TypeScript. Invoke it when your agent is scaffolding integrations: you get an architecture model from marketplace configuration through Rivet module clients to outbound API calls. Pair it with zoom-oauth when you must choose grant types and credential shapes before coding clients. The skill is procedural reference knowledge for Build-phase integration work—not a hosted MCP—so outputs are clearer module boundaries, endpoint ownership, and webhook flow diagrams your agent can implement in your repo.840installs128Setup Zoom Webhookssetup-zoom-webhooks is an operational agent skill packaged as a five-step preflight for Zoom Meeting/Webinar webhooks. It targets solo and indie builders who wire Zoom into SaaS or internal tools and keep hitting silent verification or validation failures. The skill walks through public HTTPS reachability and reverse-proxy routing, correct v0 signature construction on the untouched request body, URL-validation challenge responses, enabled event subscriptions in the Zoom app, and a respond-quick-then-process-async pattern. It is not a full Zoom SDK tutorial; it is the checklist you run when integrations look correct but events never land or signatures always fail. Use it during build when standing up or fixing Zoom listeners, and again in ship when hardening production webhook paths.840installs129Choose Zoom Approachchoose-zoom-approach is an architecture picker for solo builders adding Zoom to a SaaS or agent product. Instead of guessing SDKs from blog posts, you classify the problem—deterministic backend automation, event delivery, embedded meetings, custom video UI, in-client apps, MCP tool access, phone, or contact center—and map it to the documented Zoom skill path in the plugin family. The framework pushes the smallest correct surface first, which keeps OAuth scopes, webhook verification, and client embedding complexity proportional to what you actually ship. Use it when a PRD mentions “Zoom” without naming REST vs SDK vs MCP, when you are comparing webhooks vs WebSockets for server reactions, or when an agent needs meeting data without rebuilding Zoom’s meeting model in Video SDK. It is planning knowledge, not a credential store: you still implement auth, signing, and rate limits in the chosen follow-on skill.839installs130Setup Zoom WebsocketsSetup Zoom WebSockets is an agent skill that walks solo builders through managing live WebSocket connections to Zoom using Server-to-Server OAuth. It is for indie developers and small teams who already have Zoom API credentials and need a dependable pattern for opening sockets, streaming events, surviving disconnects, and closing sessions cleanly. The guide sequences token generation, connection open, real-time event handling, reconnect logic, and teardown so you do not ship a brittle listener that dies every hour when tokens expire. Example code uses axios and Node-style Buffer encoding for the OAuth call, which maps cleanly to backend services or agent-driven scaffolding. Use it while you are building meeting analytics, attendance bots, or workflow triggers that must react to Zoom events in near real time. It does not replace Zoom’s full platform docs for scopes, subscriptions, or compliance review—you still validate those in Validate and harden in Ship—but it compresses the repetitive connection boilerplate into copy-ready steps your coding agent can implement.839installs131Zoom Oauthzoom-oauth is a decision-oriented agent skill that maps real Zoom integration scenarios to the four supported OAuth 2.0 flows—Server-to-Server, User OAuth, Device Flow, and Chatbot credentials—so solo builders do not ship the wrong grant type to production. It clarifies when the app acts on its own account versus when a user must authorize, and it anchors the correct authorization and token URLs. Use it in Validate when you are scoping whether your SaaS serves other Zoom accounts or only automates your own tenant, then again in Build when implementing token exchange and refresh behavior. The matrix ties grant types like account_credentials and authorization_code to concrete outcomes such as one-hour S2S tokens without refresh. After the flow is fixed, hand off to rivet-sdk for module clients and webhooks. It is reference procedural knowledge, not a secrets store—your agent should output a flow choice and credential checklist, not live client secrets.839installs132Plan Zoom IntegrationPlan Zoom Integration is an agent skill that converts a Zoom integration or app concept into an actionable implementation plan for solo builders and small teams. You invoke it when you know what you want users to do in Zoom but lack a clear architecture, credential model, or delivery order. The workflow captures the target user flow and success criteria, selects the right Zoom products and supporting services, documents auth requirements and scopes, and splits work into prototype, core integration, reliability, and launch phases. It forces hard risks—OAuth misconfiguration, webhook verification, SDK environment limits, marketplace review, and MCP client constraints—onto the plan before coding. Output includes an architecture summary, required APIs, an auth checklist, phased milestones, risks, open questions, and the smallest deliverable that proves the design. It pairs naturally with Zoom OAuth setup and concrete build skills for meeting apps and bots.838installs133Build Zoom BotBuild-zoom-bot is a phase-specific agent skill for solo builders automating Zoom meetings—recorders, notetakers, or agents that react to live session data. It forces clarity on whether the bot must join, observe, transcribe, summarize, or take action, then steers implementation through Meeting SDK and RTMS for real-time paths, with REST for resource management and Webhooks for asynchronous events. The workflow emphasizes backend orchestration, storage, post-processing, and event flow rather than a single API call. It explicitly warns against treating batch transcription and live media as one pipeline and against designing UI before join authority and auth are settled. Use when integrating Zoom into agent products, internal ops bots, or meeting-centric workflows alongside the referenced Zoom plugin skills.837installs134Build Zoom Contact Center Appbuild-zoom-contact-center-app is procedural guidance for solo and indie Android builders integrating Zoom Contact Center (CCI) into Kotlin apps. It maps the full SDK lifecycle: initialize once in Application.onCreate, build a ZoomCCItem per channel (chat, video, scheduled callback), call init and listeners, then launch with the right login and fetchUI sequence for each sdkType. The skill emphasizes cleanup paths—ending engagements, logging off to stop callbacks, and releasing the service in onDestroy—so production apps do not leak sessions or duplicate listeners. Campaign mode adds a reinitialization path when users pick a channel from campaign metadata. Use it when you are past prototype and need a consistent pattern for ZoomCCInterface services, server types, and channel switches without re-reading scattered SDK notes in chat.836installs135Build Zoom Meeting Appbuild-zoom-meeting-app is an agent skill for solo builders embedding real Zoom meetings in web or mobile products. It walks you through confirming whether the user needs an actual Zoom meeting versus a custom video session, then routes implementation to the Meeting SDK when embeds must join standard Zoom meetings. The workflow pulls the right platform references, adds Zoom REST APIs only for meeting creation or resource management, and layers webhooks or RTMS only when the use case explicitly requires them. It is aimed at indie developers wiring OAuth, join tokens, and lifecycle UX who do not want to discover late that Video SDK was the wrong choice. Use it when implementing Meeting SDK joins, meeting lifecycle flows, or deciding between Meeting SDK and Video SDK before you commit architecture and sprint time.835installs136Debug Zoomdebug-zoom is an agent skill for solo builders stuck on a broken Zoom stack without reading the entire documentation set. You pass symptoms, an error string, or a failing flow; the skill classifies the failure layer, asks only for the missing evidence (platform, request/response, webhook payload, or code path), and returns two to four plausible causes ordered by likelihood. It then routes you to the right deep references under the Zoom skills tree and ends with a short verification plan so you can confirm the fix. The skill fits both late Build integration debugging and Operate incidents when OAuth, REST, webhooks, Meeting or Video SDK behavior, or MCP transport misbehaves. It is deliberately narrow: hypothesis ranking plus targeted links, not a generic chat troubleshooting loop.834installs137Build Zoom Phone Integrationbuild-zoom-phone-integration is an agent skill that teaches solo builders how to architect Zoom Phone inside a web app: Smart Embed in the browser, a CRM-side event bridge, and a backend that holds OAuth tokens and calls Zoom Phone REST APIs. It walks through provisioning, authorization, iframe readiness, engagement, persistence keyed by callId, and post-call notes or recording links. The readme emphasizes version drift strategy—normalize payloads in one adapter, feature-flag optional fields, and keep webhook and embed handlers forward-compatible. It fits indie SaaS teams adding click-to-call or SMS from a customer record without leaking credentials to the client. Use it when you already committed to Zoom Phone and need a concrete service pattern rather than ad-hoc SDK snippets.833installs138Setup Zoom McpSetup Zoom MCP is a planning skill for solo builders and agent operators who need Zoom meeting or account data inside Claude-driven workflows without blindly enabling every connector. It walks you through whether the job is deterministic automation, LLM tool orchestration, or a hybrid where REST handles what MCP cannot, then documents transport assumptions, auth scopes, and client capability limits before anyone writes integration code. You invoke it with a concrete use case argument so the agent ends with a minimal proof-of-concept sequence instead of a vague “turn on MCP” checklist. It does not replace Zoom’s official OAuth console work or production monitoring; it front-loads architecture choices that prevent rework when REST endpoints are still required for batch jobs or unsupported actions. Pair it early when evaluating Zoom in a knowledge-work plugin stack, then follow the linked design skills once the strategy is approved.833installs139Zoom Generalzoom-general is a compact reference skill for Zoom Marketplace architecture aimed at solo builders who must scope an integration before touching code. It explains when to use a flexible General App versus backend-only Server-to-Server OAuth or event-only Webhook apps, and how Admin scopes differ from User self-service authorization. The readme lists product surfaces—from Meetings and Webinars through Phone, Team Chat, and Workflows—and which embed SDKs map to each experience. It also covers Access configuration for webhook verification and optional WebSockets. Use it early to avoid wrong app type rework, then pair with surface-specific skills such as build-zoom-phone-integration when Phone is in scope. The content is taxonomy and decision support, not step-by-step implementation.833installs140Build Zoom Team Chat Appbuild-zoom-team-chat-app is an integration primer for solo builders adding Zoom Team Chat to a product, internal tool, or agent workflow. Before any code, it forces a single architectural choice: user-scoped Team Chat API messaging versus bot-scoped Chatbot API experiences. The matrix maps real jobs—pipeline notifications appearing as a person, scripted user automation, interactive bots, slash commands, and rich cards—to the API surface and credential model you must use. That upfront guardrail matters because Zoom splits OAuth patterns and URL namespaces; swapping them produces auth errors and messages in the wrong identity. The readme is guide-first rather than a full scaffold generator, so your agent should implement webhooks, token storage, and channel routing after the selection is locked. It fits backend or agent features where Zoom is the notification or collaboration channel, not a generic Zoom Meetings skill.832installs141Ui Toolkit/Webui-toolkit/web is an agent skill from Anthropic’s knowledge-work-plugins collection aimed at solo and indie builders who want structured help when implementing web user interfaces. Catalog metadata positions it as a narrow frontend capability inside a larger ui-toolkit stack, so it is best invoked once you have a product direction and are actively shaping pages, components, or layout behavior rather than during early research. Because the published stub emphasizes identity and install path more than a long procedural checklist, treat it as companion procedural knowledge that steers the agent toward consistent web UI conventions instead of a standalone code generator. It matters for Prism audiences who browse by journey phase: you install it when frontend work is the bottleneck and you already run Claude Code or similar agents with skills enabled. Expect the skill to reinforce toolkit-specific patterns rather than replace a full design system or framework documentation.832installs142Build Zoom Video Sdk AppBuild-zoom-video-sdk-app is a platform skill for indie Android builders who need a fully custom in-app video experience on Zoom’s Video SDK rather than embedding the standard Meeting UI. The documented flow starts with a backend that mints short-lived JWTs from Video SDK credentials, then has the Android client initialize the SDK, join with sessionName plus token, and mirror participant and media state from SDK events into your own UI. It stresses explicit camera, microphone, and screen-share control, chat and command channels when needed, and disciplined teardown when users leave sessions. Security guidance is clear: never ship Key/Secret in the app and never confuse Video SDK payloads with Meeting SDK join fields. Use this skill when you are in active Build work on a telehealth, education, or community app that must own layout, branding, and session logic. Expect to already run Android Studio with the mobilertc package and a working token route before the agent drafts client code.831installs143Zoom RtmsZoom-rtms is an operational agent skill that compresses Zoom Real-Time Media Streams setup into a fast preflight runbook for solo builders shipping meeting intelligence, captions, or analytics pipelines. It reinforces that RTMS is backend-first: your service ingests media after lifecycle events fire, while any frontend only subscribes to processed output. The steps force you to validate architecture assumptions, event-triggered kickoff, correct identifiers per product surface, webhook responsiveness, and connection heartbeat behavior before chasing low-level codec bugs. That ordering catches the highest-frequency handshake and duplicate-stream failures called out in the doc. Treat it as a gate during Build when wiring Zoom marketplace apps, and again in Ship or Operate when webhooks misbehave under load. It complements platform skills like Video SDK app builds by keeping stream IDs and session references consistent across products.831installs144Build Zoom Virtual AgentBuild Zoom Virtual Agent is an agent skill for solo and indie builders shipping Android apps that host Zoom Campaign experiences inside a WebView. It walks through a seven-step lifecycle: construct the launch intent with URL and policy flags, enable JavaScript (and optional multi-window), inject user context before interaction, attach the bridge when zoomCampaignSdk:ready fires, route @JavascriptInterface callbacks, handle URL actions and handoff payloads, then tear down the view and references on exit. Copy-paste Kotlin snippets cover native bridge registration, support_handoff JSON relay, and URL governance for target=_blank flows. Use it when you already chose Zoom’s campaign SDK and need a disciplined native shell rather than ad-hoc WebView hacks. It does not replace Zoom account setup, server-side campaign configuration, or iOS parity work—you still own compliance review and production monitoring.830installs145Debug Zoom IntegrationDebug Zoom Integration is a focused debugging skill for solo builders who already shipped Zoom-related code and hit failures in authentication, REST calls, webhook verification, SDK joins, real-time media, or Zoom MCP transport. It enforces a strict triage order so you isolate configuration versus request construction versus runtime behavior before proposing fixes. The agent is guided to collect exact error text, platform and SDK versions, sample payloads, and whether failures are intermittent. Reference routing sends you to the right companion skill in the knowledge-work plugin family for OAuth, APIs, webhooks, Meeting SDK, Video SDK, RTMS, or Zoom MCP specifics. Deliverables are operational: name the most likely failing layer, list ranked hypotheses, outline a minimal fix plan, and define verification steps. Use it when MCP or app dashboards show connected but meetings, tokens, or events still fail in production-like tests.830installs146Zoom Apps SdkZoom Apps SDK guidance documents how a Zoom App is structured as a web frontend loaded inside Zoom’s embedded browser and a backend that owns OAuth, token exchange, and Zoom REST API calls. Solo builders use it when they want meeting-aware features—context from `getMeetingContext`, SDK configuration, and API calls proxied through their own server—without guessing how the client WebView and Node/Express backend split responsibility. The skill emphasizes `@zoom/appssdk` as the bridge between your UI and the Zoom client, which matters for indie teams shipping marketplace or internal Zoom integrations where mis-wiring OAuth or calling APIs from the browser breaks review and security expectations. It fits the build phase for anyone combining HTML/CSS/JS in Zoom with a small API layer, and it stays phase-specific because the work is integration scaffolding rather than a journey-wide methodology.830installs147Zoom Cobrowse SdkZoom Cobrowse SDK skill walks through how to add co-browsing to a solo builder’s support or onboarding flow: pick CDN or npm integration, generate JWTs only on the server, and run the customer/agent two-role pattern with PIN-based session join. It stresses short-lived tokens, distinct credentials per role, and event-driven session states so you do not leak SDK secrets into the browser. For indie SaaS teams embedding Cobrowse in their own agent console or customer page, the skill orders the objects you usually create—session records, PINs, expiry—and points to official authorization and get-started references. It is build-phase and integration-focused because the outcome is working Cobrowse plumbing, not marketing distribution or production incident response, though the same integration often supports Grow support workflows once shipped.830installs148Interactive Dashboard BuilderInteractive Dashboard Builder is an agent skill from Anthropic’s knowledge-work plugins that helps solo builders and small teams produce polished, serverless HTML dashboards. It encodes a repeatable page skeleton—viewport meta, pinned Chart.js CDNs, dashboard container, header filters, KPI row, and chart sections—so you can drop in datasets and ship a file anyone can open in a browser. Use it when a spreadsheet export is not enough but standing up Grafana or a full React app is overkill. The skill emphasizes interactivity through client-side filters and Chart.js configurations rather than API wiring. It spans Validate when you prototype metrics storytelling, Build when you harden the layout, and Grow when you refresh lifecycle or analytics narratives for stakeholders.596installs