agent-skills plugin
A growing set of Claude Code skills in one plugin.
How to install the agent-skills plugin
Run both commands inside Claude Code. The first adds the marketplace once; the second installs the plugin.
/plugin marketplace add MarvelCollin/agent-skills
/plugin install agent-skills@marvelcollinWhat's inside agent-skills
7 parts read from the plugin's own files.
Skills 7
Claude loads it when the task matches- backend
Build, review and load test backend services to an expert standard. When writing server code, applies hard rules B1 to B18 (authentication and authorization on every route, object and field, the server decides roles and prices, no secrets or sessions in the browser, RLS and private buckets for data the client can reach, validated input, no N+1 and bounded queries, indexes and constraints, transactions and idempotency, timeouts and rate limits, safe logs and errors, metrics and health checks, background jobs, validated config, correct money and time types, tests). Reviews an existing backend with a static scan and a scored report, and runs smoke, load, stress, spike, soak and breakpoint tests on a service the user owns to find the bottleneck.
- checkprod
Check whether a web app is really ready for production. Grades 40 launch items in two groups and backs every pass with proof. Frontend FE1 to FE20 covers the custom 404, a CTA above the fold, a title and description per page, the Open Graph image, the favicon set, robots.txt, sitemap, alt text, mobile breakpoints, loading, error and success states, privacy and terms pages, cookie consent, analytics, real contact details with no placeholder content, and HTTPS on the real domain. Backend BE1 to BE20 covers hidden API keys, RLS, admin routes, rate limits, storage buckets, input validation, auth sessions, secret history and dependency scans, trimmed responses and production config. Runs a static project scan and a live URL check, hands active attack tests to /security, then fixes what failed.
- explain
Explain anything the user does not understand in the simplest true way, so it is easy to understand and easy to remember. Gives the core idea in one plain sentence, one everyday picture with its limit stated, a concrete example before any definition, at most three new ideas, the common wrong idea named, and one small question that makes the idea stick. Plain words, short sentences, every needed term defined in plain words, no filler. Built on learning science (cognitive load, worked examples, concrete before abstract, analogy research, fixing misconceptions, retrieval practice, self-explanation) and the methods of great explainers. Also teaches a topic step by step, quizzes the user, checks the user's own explanation for gaps, or helps the user memorize something (a template, formula, list or steps) with a logic map, fading cues, recall drills and a spacing schedule.
- research
Check, audit, draft and review research documents (papers, theses, reports, proposals) in any field, in LaTeX, Markdown, Word text or plain text. Verifies references against registries so every citation is real, resolvable, matches its source and sits where a claim needs it. Audits logic and data in a fixed order (claims against evidence, numbers across abstract, text and tables, missing baselines, definitions and limitations, causal overreach, circular reasoning, contradictions between sections), quoting the text and location of every finding. Drafts or restructures sections to the field's conventions, and simulates a tough peer reviewer with scores, ranked objections and a rebuttal plan. Never invents a reference, DOI, author or number.
- revision-doc
Build a before and after revision report as a Word document for a batch of UI or product changes. Captures a before screenshot of every screen before the code changes and an after screenshot of the same screen at the same size, draws red boxes on what changed, and writes a landscape .docx with one page per change holding the problem behind each request, what changed and how it was checked, with the before and after shots side by side. The document is authored by the person signed in to the computer, not python-docx, and carries a real preview thumbnail so File Explorer and Finder show the first change instead of a blank white page.
- security
Run an authorized security review of a web app you own or are cleared to test, especially local and dev builds. Dev mode finds the running dev server, writes the scope itself and screens the build end to end with no setup questions. Works in phases that confirm authorization and scope, map the attack surface, then test for common vulnerability classes (injection, broken auth and access control, SSRF and server-side, XSS and client-side, session handling, business logic, misconfiguration, exposed secrets, secrets and tokens visible in the browser, client-controlled values like role and price, tables without RLS and public storage buckets, API and LLM issues) against the OWASP WSTG and ASVS, verify each finding with evidence to kill false positives, then write a report with severity and remediation. Can also do a static secure-code review without running anything.
- uiux
Build and audit web UI with real UX discipline. When building or changing UI, applies hard rules R1 to R20 (a random Color Hunt palette with no purple unless the user picks colors, no icon tile stat cards, icons in rounded squares or pill badges, search and filters on every table column, overlays that size to content, form modals with pinned actions and one scroll area, custom controls and scrollbars instead of browser built-ins, specific and consistent layouts, no semicolons or em dashes in copy, backend pagination with skeletons, real avatars, the project's conventions and file layering, no AI template tells like sparkle icons or fake testimonials, launch ready pages with a CTA above the fold, meta, alt text and a 404, forms on one aligned grid with one control height, short calm motion on transform and opacity that never janks, screens that fit before scrolling with sticky headers and tabs, and a polish pass for radius, headlines, numbers and alignment) plus navigation, color and the Eight Golden Rules, fo
Using agent-skills in Claude Code
Skills need no command. Claude reads them when your request matches what they describe.
Update, turn off or remove
- Update the plugin (shell)
- claude plugin update agent-skills@marvelcollin
- Refresh the marketplace
- /plugin marketplace update marvelcollin
- Turn it off for now
- /plugin disable agent-skills@marvelcollin
- Remove it
- /plugin uninstall agent-skills@marvelcollin
Skills in agent-skills
- backendMarvelCollin/agent-skills
Build, review and load test backend services to an expert standard. When writing server code, applies hard rules B1 to B18 (authentication and authorization on every route, object and field, the server decides roles and prices, no secrets or sessions in the browser, RLS and private buckets for data the client can reach, validated input, no N+1 and bounded queries, indexes and constraints, transactions and idempotency, timeouts and rate limits, safe logs and errors, metrics and health checks, background jobs, validated config, correct money and time types, tests). Reviews an existing backend with a static scan and a scored report, and runs smoke, load, stress, spike, soak and breakpoint tests on a service the user owns to find the bottleneck.
38 files5stars - checkprodMarvelCollin/agent-skills
Check whether a web app is really ready for production. Grades 40 launch items in two groups and backs every pass with proof. Frontend FE1 to FE20 covers the custom 404, a CTA above the fold, a title and description per page, the Open Graph image, the favicon set, robots.txt, sitemap, alt text, mobile breakpoints, loading, error and success states, privacy and terms pages, cookie consent, analytics, real contact details with no placeholder content, and HTTPS on the real domain. Backend BE1 to BE20 covers hidden API keys, RLS, admin routes, rate limits, storage buckets, input validation, auth sessions, secret history and dependency scans, trimmed responses and production config. Runs a static project scan and a live URL check, hands active attack tests to /security, then fixes what failed.
7 files5stars - explainMarvelCollin/agent-skills
Explain anything the user does not understand in the simplest true way, so it is easy to understand and easy to remember. Gives the core idea in one plain sentence, one everyday picture with its limit stated, a concrete example before any definition, at most three new ideas, the common wrong idea named, and one small question that makes the idea stick. Plain words, short sentences, every needed term defined in plain words, no filler. Built on learning science (cognitive load, worked examples, concrete before abstract, analogy research, fixing misconceptions, retrieval practice, self-explanation) and the methods of great explainers. Also teaches a topic step by step, quizzes the user, checks the user's own explanation for gaps, or helps the user memorize something (a template, formula, list or steps) with a logic map, fading cues, recall drills and a spacing schedule.
7 files5stars - researchMarvelCollin/agent-skills
Check, audit, draft and review research documents (papers, theses, reports, proposals) in any field, in LaTeX, Markdown, Word text or plain text. Verifies references against registries so every citation is real, resolvable, matches its source and sits where a claim needs it. Audits logic and data in a fixed order (claims against evidence, numbers across abstract, text and tables, missing baselines, definitions and limitations, causal overreach, circular reasoning, contradictions between sections), quoting the text and location of every finding. Drafts or restructures sections to the field's conventions, and simulates a tough peer reviewer with scores, ranked objections and a rebuttal plan. Never invents a reference, DOI, author or number.
11 files5stars - revision-docMarvelCollin/agent-skills
Build a before and after revision report as a Word document for a batch of UI or product changes. Captures a before screenshot of every screen before the code changes and an after screenshot of the same screen at the same size, draws red boxes on what changed, and writes a landscape .docx with one page per change holding the problem behind each request, what changed and how it was checked, with the before and after shots side by side. The document is authored by the person signed in to the computer, not python-docx, and carries a real preview thumbnail so File Explorer and Finder show the first change instead of a blank white page.
6 files5stars - securityMarvelCollin/agent-skills
Run an authorized security review of a web app you own or are cleared to test, especially local and dev builds. Dev mode finds the running dev server, writes the scope itself and screens the build end to end with no setup questions. Works in phases that confirm authorization and scope, map the attack surface, then test for common vulnerability classes (injection, broken auth and access control, SSRF and server-side, XSS and client-side, session handling, business logic, misconfiguration, exposed secrets, secrets and tokens visible in the browser, client-controlled values like role and price, tables without RLS and public storage buckets, API and LLM issues) against the OWASP WSTG and ASVS, verify each finding with evidence to kill false positives, then write a report with severity and remediation. Can also do a static secure-code review without running anything.
27 files5stars - uiuxMarvelCollin/agent-skills
Build and audit web UI with real UX discipline. When building or changing UI, applies hard rules R1 to R20 (a random Color Hunt palette with no purple unless the user picks colors, no icon tile stat cards, icons in rounded squares or pill badges, search and filters on every table column, overlays that size to content, form modals with pinned actions and one scroll area, custom controls and scrollbars instead of browser built-ins, specific and consistent layouts, no semicolons or em dashes in copy, backend pagination with skeletons, real avatars, the project's conventions and file layering, no AI template tells like sparkle icons or fake testimonials, launch ready pages with a CTA above the fold, meta, alt text and a 404, forms on one aligned grid with one control height, short calm motion on transform and opacity that never janks, screens that fit before scrolling with sticky headers and tabs, and a polish pass for radius, headlines, numbers and alignment) plus navigation, color and the Eight Golden Rules, fo
56 files5stars
agent-skills plugin questions
How do I install the agent-skills plugin?
In Claude Code run /plugin marketplace add MarvelCollin/agent-skills, then /plugin install agent-skills@marvelcollin.
What does the agent-skills plugin include?
agent-skills bundles 7 skills.
How do I update or remove the agent-skills plugin?
Update it with claude plugin update agent-skills@marvelcollin in your shell, or from the Installed tab of /plugin. Remove it with /plugin uninstall agent-skills@marvelcollin.
What license is the agent-skills plugin under?
MIT, as stated in the plugin's manifest.