alex-macra
UserReusable agent skills, deterministic activation, safety hooks, and installers for Claude Code and OpenAI Codex.
Categories
Indexed Skills (32)
fast-pr-workflow
Git and GitHub PR workflow: branch from the task, commit, and create or update the pull request without merging or rebasing unless asked. Use to create PR, update PR, commit, push, or ship this work. Merge, squash, and land requests route here so the main-branch guard applies.
e2e-qa
End-to-end testing with Playwright and Cypress plus real CLI and HTTP journeys: test architecture, locator discipline, network interception, flake elimination, CI parallelization. Use when writing, fixing, or debugging E2E, browser, or user-journey tests. Unit and integration testing belongs to qa-automation.
qa-automation
Writing and running automated tests: unit and integration design, fixtures, mocks vs real dependencies, coverage, flake hunting, CI integration. Use when adding, running, fixing, or debugging tests with pytest, vitest, jest, or mocha. Browser, CLI, and HTTP end-to-end flows belong to e2e-qa.
see-it-live
Launch the app and confirm the specific change is actually visible and working: screenshot the UI, curl the endpoint, or run the command. Doubles as the fast smoke pass: boots, core path works, key endpoint responds. Use for run the app, start the server, preview it, see it live, smoke test, sanity check, or post-deploy check.
e2e-qa
End-to-end testing with Playwright and Cypress plus real CLI and HTTP journeys: test architecture, locator discipline, network interception, flake elimination, CI parallelization. Use when writing, fixing, or debugging E2E, browser, or user-journey tests. Unit and integration testing belongs to qa-automation.
qa-automation
Writing and running automated tests: unit and integration design, fixtures, mocks vs real dependencies, coverage, flake hunting, CI integration. Use when adding, running, fixing, or debugging tests with pytest, vitest, jest, or mocha. Browser, CLI, and HTTP end-to-end flows belong to e2e-qa.
see-it-live
Launch the app and confirm the specific change is actually visible and working: screenshot the UI, curl the endpoint, boot the scene, or run the command. Use for run the app, launch the app, start the server, spin up, preview it, see it live, does it render, or before declaring a user-visible change done.
smoke-test
Fast critical-path sanity check: app boots, core path works, key endpoint responds. Run after integrating a change and again after deploy or merge, not as a full user journey. Use for smoke test, sanity check, critical path verification, does it still work, or a post-deploy check.
code-reuse
Reuse before writing: find the existing function, util, hook, or component and extend it instead of adding a near-duplicate. Use when adding a helper or utility, consolidating duplicate or copy-paste code, or reviewing a diff for reuse.
spark-task-planning
Turn approved architecture and repository evidence into dependency-ordered implementation packets for Qwen, Qwen3-Coder-Next, or DGX Spark. Use only when the request explicitly targets one of those execution contexts; generic task packets and Apache Spark planning do not belong here.
spark-task-research
Repository-grounded research for DGX Spark and Qwen3-Coder-Next implementation work: verify files, symbols, versions, dependencies, commands, and risks before task planning. Use for DGX Spark task research, Qwen task research, or preparing repository evidence for implementation packets. Task decomposition belongs to spark-task-planning.
delivery-loop
Execute an approved Markdown task packet through a bounded loop: validate it against the repository, implement it, run targeted tests and builds, fix failures, and retest. Use for delivery loops, executing task Markdown, implementing a task packet, or test-build-fix-retest work. Qwen or DGX research and planning belong to the Spark task skills; PR work belongs to fast-pr-workflow.
a11y-audit
Accessibility audit of a page or flow: axe-core, Lighthouse, and a WCAG 2.2 AA manual checklist, severity-ranked with fixes. Use for an a11y audit, WCAG conformance, screen reader or keyboard testing, or contrast ratios. Building accessible components belongs to web-dev-frontend.
adversarial-review
Independent second-pass review that tries to break the change: invariants, edge cases, error paths, concurrency, ending in a concrete failing input and a verdict. Use for an adversarial review, red team pass, second opinion, try to break this, or poke holes. Deep security passes belong to security-review.
architect-review
Clean code and architecture evaluation: module boundaries, coupling and cohesion, abstraction levels, naming, dependency direction, complexity hotspots, refactoring strategy. Use for a code, design, or architecture review, refactor planning, tech debt evaluation, or is this design right.
automation
Repeatable infrastructure: CI/CD pipelines, GitHub Actions, deploy and release scripts, dev-server orchestration, cron and scheduled jobs, container builds, backups, rollback. Use when setting up or fixing CI, writing a Dockerfile or deploy script, or on any deploy, release, pipeline, or schedule request.
code-comments
When and how to write code comments: the default is none, since code and tests carry the meaning and only the non-obvious WHY earns one. Use when adding or removing comments, writing docstrings, JSDoc, or TSDoc, deciding whether code needs documenting, or reviewing a comment-heavy diff.
security-review
Security review of a codebase or a target you are authorized to test: OWASP Top 10, secrets, auth, input validation, supply chain, infra hygiene. Use for a security audit, vulnerability or CVE check, pentest-style review, hardening, secret-leak hunts, or is-this-secure questions.
task-research
Research a task before building: codebase prior art first, then docs and web, separating facts from assumptions, ending in a short findings brief that feeds a plan or ticket. Use for research, investigate, look into, find out how, what's the best way, compare approaches, spikes, or feasibility checks.
web-dev-backend
Web backend work in Node/TypeScript (Express, Fastify) and Python (FastAPI, Django): routes, middleware, async I/O, validation, type discipline. Use when building or reviewing backend services, API endpoints, database schemas and migrations, or files under api/, backend/, or server/.
web-dev-frontend
Web frontend in TypeScript with React, Vue, or Svelte: components, hooks, state management, styling and design tokens, accessibility, browser APIs, performance. Use when building or changing components, pages, forms, or UI state, theming, or any *.tsx, *.vue, *.svelte, or frontend/** file.
a11y-audit
Accessibility audit pass over a page or flow: axe-core, Lighthouse, and a WCAG 2.2 AA manual checklist, reported severity-ranked with fixes. Use for an accessibility or a11y audit, WCAG conformance, screen reader or keyboard navigation testing, or contrast ratios. Building accessible components belongs to web-dev-frontend.
adversarial-review
Independent second-pass review that tries to break the change: invariants, edge cases, error paths, security, concurrency, ending in a concrete failing input and a verdict. Use for an adversarial review, red team pass, second opinion, devil's advocate check, try to break this, poke holes, or what could go wrong.
architect-review
Clean code and architecture evaluation: module boundaries, coupling and cohesion, abstraction levels, naming, dependency direction, complexity hotspots, refactoring strategy. Use for a code, design, or architecture review, refactor planning, tech debt evaluation, or is this design right.
automation
Repeatable infrastructure: CI/CD pipelines, GitHub Actions, deploy and release scripts, dev-server orchestration, cron and scheduled jobs, container builds, backups, rollback. Use when setting up or fixing CI, writing a Dockerfile or deploy script, or on any deploy, release, pipeline, or schedule request.
code-comments
When and how to write code comments: the default is none, since code and tests carry the meaning and only the non-obvious WHY earns one. Use when adding or removing comments, writing docstrings, JSDoc, or TSDoc, deciding whether code needs documenting, or reviewing a comment-heavy diff.
delivery-loop
Bounded task-delivery conductor: research, plan, implement, quality checks, live verification, PR readiness, code review, adversarial review, smoke test, merge handoff. Use for run the delivery loop, quality loop, review loop, work this in a loop, run the orchestrator, or any guarded end-to-end workflow.
fast-pr-workflow
Git and GitHub PR workflow: check the branch, branch from the task name, stage, commit, and create or update the pull request without merging or rebasing unless asked. Use to create PR, update PR, commit changes, push branch, or ship this work. Merge, squash, and land requests route here so the protected-branch policy applies.
security-review
Security posture review of a codebase or a running target you are authorized to test: OWASP Top 10, secrets and leaked credentials, auth, input validation, supply chain, transport, infra hygiene. Use for a security audit, vulnerability or CVE check, authorized pentest-style review, hardening, secret-leak hunts, or is-this-secure questions.
task-research
Research a task before building: codebase prior art first, then docs and web, separating facts from assumptions, ending in a short findings brief that feeds a plan or ticket. Use for research, investigate, look into, find out how, what's the best way, compare approaches, spikes, or feasibility checks.
web-dev-backend
Web backend work in Node and TypeScript (Express, Fastify) and Python (FastAPI, Django, Flask): routes, middleware, async I/O, validation, data pipelines, type discipline, dependency hygiene. Use when building or reviewing backend services, API endpoints, database schemas and migrations, or files under api/, backend/, or server/.
web-dev-frontend
Web frontend in TypeScript with React, Vue, or Svelte: components, hooks, state management, styling and design tokens, accessibility, browser APIs, performance. Use when building or changing components, pages, forms, or UI state, theming, or any *.tsx, *.vue, *.svelte, or frontend/** file.
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.