The core pack (agents-pack-core 0.31.1) includes 35 skills . Coding agents usually select a skill from a natural-language request. To invoke one explicitly, include its ID:
Use ap-debug to investigate why this test fails intermittently.
Required management and project-memory skills are always installed. Other skills depend on your component selection.
Skill
ID
Selection
Summary
Manage Agents Pack ap-manage-agents-packrequired Manage official and user-owned cross-agent components through the safe lifecycle CLI.
Create a new cross-agent skill ap-create-new-skillrequired Create or update one canonical user-owned skill and synchronize it across selected agents.
Add an MCP server ap-add-mcprequired Add, inspect, and remove user-level remote MCP servers across coding agents.
Skill
ID
Selection
Summary
Run a design studio ap-design-studiooptional Coordinate original, isolated static HTML concepts, fresh critiques, polish, and shareable files before explicit product integration.
Explore design directions ap-explore-design-directionsoptional Generate binding creative packets with varied originality methods for a taste checkpoint or a requested batch of static prototypes.
Implement a new design ap-implement-new-designoptional Build a chosen visual direction, refine it through bounded fresh-context critiques, and verify the final interface.
Polish a design ap-design-polishoptional Remove clutter and generic visual treatments while preserving design identity, usability, and working behavior.
Skill
ID
Selection
Summary
Develop with the Vercel AI SDK ap-develop-with-vercel-ai-sdkoptional Use current AI SDK APIs, architecture, tool safety, streaming, persistence, and testing patterns.
Skill
ID
Selection
Summary
Design data models ap-design-data-modelsoptional Design maintainable relational, document, or distributed data models and migrations.
Develop APIs ap-develop-apisoptional Build thin, secure, documented API layers over reusable business and data logic.
Write database queries ap-write-database-queriesoptional Write correct, secure, performant queries with sound transaction and indexing choices.
Skill
ID
Selection
Summary
Compress project TODOs ap-compress-todosrecommended Clean up completed history while preserving active work, decisions, and priority.
Refresh repository documentation ap-refresh-repo-docsrecommended Maintain and reconcile project documentation with implementation and product intent.
Skill
ID
Selection
Summary
Handle errors reliably ap-handle-errors-reliablyrecommended Design explicit error boundaries, cleanup, retries, timeouts, and observable failures.
Validate trust boundaries ap-validate-trust-boundariesrecommended Validate untrusted files, text, and structured input at system boundaries.
Skill
ID
Selection
Summary
Design distinctive frontends ap-frontend-designoptional Create intentional, production-grade interfaces and maintain the project design system.
Review frontend quality ap-frontend-reviewoptional Review frontend implementation, usability, accessibility, responsiveness, and visual coherence.
Apply React best practices ap-react-best-practicesoptional Apply Vercel’s performance guidance for React and Next.js using version-aware, evidence-based changes.
Apply React composition patterns ap-react-composition-patternsoptional Design scalable React component APIs with composition, explicit variants, and clear state boundaries.
Skill
ID
Selection
Summary
Run a security audit ap-security-auditoptional Inspect concrete attack surfaces, validate findings, and report actionable security risks.
Skill
ID
Selection
Summary
Test a web application ap-test-web-appoptional Test real browser flows, edge cases, responsiveness, and regressions with evidence.
Skill
ID
Selection
Summary
Debug systematically ap-debugrecommended Trace reproducible failures to evidence-backed root causes and verify fixes.
Start a development session ap-start-dev-sessionrecommended Orient a new session and isolate mutable local work in a dedicated Git worktree.
Prepare to clear development context ap-clear-dev-contextrecommended Create a detailed continuation handoff before clearing a long agent context.
Continue a development session ap-continue-dev-sessionrecommended Resume work from a continuation handoff and persisted project context.
Review a plan ap-review-planrecommended Challenge a plan for mistakes, omissions, assumptions, risks, and simpler alternatives.
Run subagent-driven development ap-subagent-driven-developmentoptional Execute an implementation plan through fresh task implementers, independent reviews, repair loops, and a final branch review.
Skill
ID
Selection
Summary
Audit SEO ap-audit-seooptional Audit technical, structural, content, and measurement aspects of search visibility.
Audit GEO and AEO ap-audit-geooptional Audit AI-search eligibility, usefulness, citations, and measurement.
Skill
ID
Selection
Summary
Run market research ap-run-market-researchoptional Produce current, source-grounded market research with explicit evidence limits.
agents-pack install ap-frontend-design --dry-run
agents-pack install ap-frontend-design --yes
Or ask your coding agent to use ap-manage-agents-pack to preview and install it.
