atlas

MANUAL skill covering the full atlas lifecycle outside of task work: onboard (scaffold the docs/ SSOT plus .atlas/ internal state, inventory skills, recommend what to run next), install (verify and wire claude-mem, context-mode, hooks, project config), connectors (guided vendor MCP connector setup across domain plugins), and repair (fix a broken atlas install: marketplace, rollbacks, hooks, assets). Run with no args for onboarding plus recommendations; run with --fix to auto-repair.

Atlas architect: boot the workspace. Verify claude-mem and context-mode, scan the project, recommend tooling (confirm first), wire hooks, seed the docs/ SSOT.

This skill should be used when the user asks to "audit my codebase against The Pragmatic Programmer", run a "pragmatic audit", "check what pragmatic principles are implemented", "review codebase for pragmatic practices", "find what's missing from the pragmatic programmer in this project", or wants a structured review of a codebase/project/monorepo against the book's principles, tips, and practices. Produces a ranked report of what is implemented, partially implemented, and missing, each finding cited to the book concept and to file:line evidence. Supports --chapters to scope dimensions and --report to set output path.

This skill should be used when the user is "designing a module", "deciding an architecture", "debugging a bug", "refactoring code", "writing tests", "naming things", "handling concurrency", "decoupling components", "choosing inheritance vs composition", "adding error handling", asks "what does The Pragmatic Programmer say about X", "is this DRY", "should I use inheritance or composition", "how should I structure this", or otherwise wants the relevant Pragmatic Programmer principles, tips, or practices surfaced for the work in progress. Surfaces 1-4 relevant principles with book tip numbers and citations to the concept files.

Create or modify a reusable component that survives latency, cancellation, and partial failure (progress modal, upload widget, job panel), handling every backend state with all six interaction states rendered.

Read-only database audit via parallel subagents: inventory a live schema, reconcile it against the code, and check privileges and naming before any change.

Reproducible bug, exception, stack trace, or bad output: root-cause fix with evidence, not a patch over. Use when you want the actual cause fixed, not the symptom hidden.

Build or refactor screens, flows, or components on a single design system (shadcn/ui + Tailwind + Radix) with every state handled and verified live in the browser.

Generate a zero-trust, deny-by-default .gitignore for a named stack: allowlist intended paths, re-exclude secrets last. Use when starting or hardening a repo.

Write an idempotent endpoint remediation script using a CHECK/SET/VERIFY pattern for RMM/MDM, proving whether it changed state or was already compliant.

Orchestrate any multi-step, multi-surface, or whole-codebase engineering task (build, fix, audit, refactor, investigate) through subagents with real execution and independent verification instead of inline work, keeping docs/ the single source of truth. Triggers on orchestrate, whole-repo work, cross-layer (frontend/backend/database) bugs, and audits. To first install and configure atlas itself, use atlas-setup.

Rewrite a vague coding request into a structured, environment-aware prompt an AI agent can execute; asks up to three questions first when scope is ambiguous.

Generate an onboarding-grade README.md by inspecting the actual repo, every claim traced to a real file. Use when a repo has no README or its README is stale.

Refactor, rename, or restructure code without changing observable behavior. Use when code works but is messy, hard to navigate, or carries dead weight, and you need behavior preserved with before/after evidence.

Validate and verify a Claude Code plugin is done: a completion check that audits structure, manifest, and content quality with file:line findings and pass/fail per check, without auto-fixing.

Generates and refreshes the docs/wiki/ diagrams from docs/architecture/ by invoking the graphify skill. Keeps the wiki fresh as the codebase changes. Use when architecture docs are updated, before completion, or when wiki diagrams are stale or missing.

Match a recurring or iterative task to the best-fit reusable loop from a curated loop-library and instantiate it, handing interval/self-paced loops to the built-in /loop skill and running fan-out loops as a parallel adversarial-verify Workflow. Use when the user wants to run something repeatedly, poll for status, iterate until a condition is met, sweep a backlog, or stand up a recurring/scheduled workflow.

Audits, three modes. CODE (default): comprehensive, discovery-first code-quality and security audit of a whole codebase - correctness, OWASP/security, SOLID/DRY/KISS, risk hotspots, dead code, coverage gaps, code-vs-docs drift - as a Workflow that builds a knowledge graph, targets the hottest nodes, fans out one reviewer per dimension, and adversarially verifies every finding. ARCHITECTURE: map the codebase into feature flowcharts, find duplication, propose the simplest unified architecture. SELF: measure atlas run health, context/asset waste, and session-transcript forensics.

Implement a feature end to end when the user says "implement a feature", "add a feature that", or "build new functionality" that spans UI, API, and data. Ships working with verified evidence, not 'should work'. Dispatches the atlas squad in parallel and closes with an independent verifier.

Produce a dense session handoff so a fresh session resumes with zero re-discovery. Use at a checkpoint before context fills, a break, or handing off work.

Launch a remediation session preloaded with a finding from the latest audit hub; use after an atlas-audit run. No args lists findings.

Organizational configuration layer for the atlas plugin. When an organization deploys atlas, this skill ensures every role and department is provisioned with the right skills, agents, connectors, branding, and compliance policies. Use when setting up an org's atlas deployment, onboarding a new department, enforcing org branding on coding-agent outputs, or routing a user to their department's toolset. Triggers on org setup, department onboarding, brand enforcement, policy compliance, and connector provisioning.

Use when asked to run a UX test swarm, full UI/UX test pass, persona testing review, or pre-release frontend sweep on any web app, or to re-test after fixes from a previous run. Discovers the target app from the repo automatically and adapts to any frontend stack. Drives persona generation, scripted data entry, real-browser walkthroughs, fuzzing, and an independent calc oracle, then gates on whether the CLIENT surface is actually correct.

Generate or improve a company-specific data analysis skill by extracting tribal knowledge from analysts. Bootstrap a new data-analysis skill from a warehouse schema, or iterate on an existing skill by adding domain reference files. Use when an analyst wants Claude to understand their company's specific data warehouse, terminology, metrics definitions, and common query patterns.

Profile and explore datasets to understand their shape, quality, and patterns before analysis. Use when encountering a new dataset, assessing data quality, discovering column distributions, identifying nulls and outliers, or deciding which dimensions to analyze.

QA an analysis before sharing with stakeholders - methodology checks, accuracy verification, and bias detection. Use when reviewing an analysis for errors, checking for survivorship bias, validating aggregation logic, or preparing documentation for reproducibility.

Create effective data visualizations with Python (matplotlib, seaborn, plotly). Use when building charts, choosing the right chart type for a dataset, creating publication-quality figures, or applying design principles like accessibility and color theory.

Build self-contained interactive HTML dashboards with Chart.js, dropdown filters, and professional styling. Use when creating dashboards, building interactive reports, or generating shareable HTML files with charts and filters that work without a server.

Write correct, performant SQL across all major data warehouse dialects (Snowflake, BigQuery, Databricks, PostgreSQL, etc.). Use when writing queries, optimizing slow SQL, translating between dialects, or building complex analytical queries with CTEs, window functions, or aggregations.

Apply statistical methods including descriptive stats, trend analysis, outlier detection, and hypothesis testing. Use when analyzing distributions, testing for significance, detecting anomalies, computing correlations, or interpreting statistical results.

Audit designs and code for WCAG 2.1 AA compliance. Trigger with "is this accessible", "accessibility check", "WCAG audit", "can screen readers use this", "color contrast", or when the user asks about making designs or code accessible to all users.

Refactor and reorganize codebases using SOLID principles, clean architecture, strict naming conventions, and defensive programming. Trigger with "refactor this codebase", "reorganize the code", "clean up the project structure", "fix the naming", "audit the codebase", "code organization", or when someone needs help restructuring, renaming, deduplicating, or improving code quality across a project.

Create comprehensive developer handoff documentation from designs. Trigger with "handoff to engineering", "developer specs", "implementation notes", "design specs for developers", or when a design needs to be translated into detailed implementation guidance.

Manage design tokens, component libraries, and pattern documentation. Trigger with "design system", "component library", "design tokens", "style guide", or when the user asks about maintaining consistency across designs.

Generate comprehensive, production-ready README.md documentation from a full codebase analysis. Trigger with "generate readme", "create readme", "update readme", "document this project", "write documentation for this repo", or when someone needs a complete project README created or refreshed from the actual implementation.

Modernize frontend codebases into production-grade interfaces using shadcn/ui, Tailwind CSS, and Radix UI. Covers design tokens, component architecture, loading/progress feedback patterns, accessibility, dark mode, and responsive design. Trigger with "modernize the UI", "ui refactor", "migrate to shadcn", "fix the design system", "add loading states", "progress feedback", "make it look modern", or when someone needs help with frontend component architecture, design tokens, or async feedback patterns.

Plan, conduct, and synthesize user research. Trigger with "user research plan", "interview guide", "usability test", "survey design", "research questions", or when the user needs help with any aspect of understanding their users through research.

Write effective microcopy for user interfaces. Trigger with "write copy for", "help with UX copy", "what should this button say", "error message for", "empty state copy", or when the user needs help with any interface text.

Run a fallow static analysis pass on a JavaScript or TypeScript repo and synthesize a prioritized, ranked code-quality report covering changed-code risk, complexity hotspots, and architecture boundary violations. Use when the user asks "run a code quality sweep", "analyze code quality with fallow", or "what is risky in this changed code". Read-only and advisory.

Review code for bugs, security vulnerabilities, performance issues, and maintainability. Use when the user says "review this code", "check this PR", "look at this diff", or "is this code safe?", or shares code and asks for feedback.