By w159
Multi-agent coding architect with automatic self-improvement. 22 plainly named skills: the atlas architect skill (bootstrap: verify claude-mem/context-mode, scan the project, recommend tooling, wire hooks, seed docs/ SSOT), the atlas-orchestrate engine, atlas-audit (code/security audit, architecture map, atlas self-telemetry), atlas-setup (onboarding, install, connectors, repair - one of two manual skills, alongside the atlas architect skill), atlas-loop (recurring/iterative work), atlas-ux-test (UX runtime swarm), and 16 task skills (atlas-feature, atlas-debug, atlas-refactor, atlas-frontend, atlas-component, atlas-db-audit, atlas-gitignore, atlas-handoff, atlas-harden, atlas-launch, atlas-m365, atlas-prompt, atlas-readme, atlas-validate, atlas-vendor-assessment, atlas-wiki). 12 core agents, 11 hooks (session boot, prompt optimizer, bash advisor, format-after-edit, dispatch tripwire, completion gate, memory capture, auto-skill, self-improvement nudge, session-transcript ingest, atlas_doctor session-health check), and the atlas-orchestrate method: independent verification with runtime evidence before any done claim. Self-improvement: persistent memory at ~/.atlas/memory/, automatic skill creation from session transcripts at ~/.claude/skills/, skill lifecycle curation, and context optimization that disables unused skills/agents. Org deployment (11 departments) lives in the separate armada plugin.
Pre-done completeness auditor. Hunts unverified claims, unread sources, unexercised paths, unmet requirements; returns a gap list and refutes 'done' on a load-bearing gap. Defers docs-drift to docs-auditor. Never fixes.
Read-only database prober. Inspects SQL/Postgres schema, RLS policies, runtime-role GRANTs, indexes, constraints, and EXPLAIN plans. Read-only: no writes or migrations, only proposals. Returns findings with evidence.
READ-ONLY drift auditor for the canonical atlas project structure (docs-ssot.md). Compares docs/ (CHANGELOG, ROADMAP, architecture, AGENTS.md), the .atlas/ structure, root entry files, and .gitignore against real code and returns a per-area verdict (current/stale/missing) with file:line evidence. Never writes.
Post-ship maintainer and enforcer of the canonical atlas project structure (docs-ssot.md). Writable scope: docs/**, the durable .atlas/ subfolders (findings/, audits/, decisions/, archive/), the root entry files (README.md, AGENTS.md, CLAUDE.md), and .gitignore. Never edits source code. Updates CHANGELOG and ROADMAP (moving verified done items to CHANGELOG with date + evidence), distills verified findings.json entries into the dated .atlas/findings/ ledger, maintains docs/wiki/, keeps .gitignore zero-trust and current, and flags missing canonical structure for atlas-setup.
Read-only codebase explorer. Maps a feature, module, or call path, locates the symbol owning a behavior, or summarizes structure without reading whole files. Returns a compact map with file:line refs, not dumps.
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.
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.
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.
Admin access level
Server config contains admin-level keywords
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimnpx claudepluginhub w159/atlas --plugin atlasBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
This plugin requires configuration values that are prompted when the plugin is enabled. Sensitive values are stored in your system keychain.
auvik_regionAuvik shard: us1 (default), us2, us3, us4, eu1, eu2, au1, ca1. Leave blank for us1.
${user_config.auvik_region}cipp_api_keyLegacy static bearer token. Use this OR the OAuth trio below.
${user_config.cipp_api_key}auvik_api_keyAuvik portal API key.
${user_config.auvik_api_key}cipp_base_urlURL of your deployed CIPP Azure Function (required for the CIPP connector). Leave blank to disable CIPP.
${user_config.cipp_base_url}auvik_usernameAuvik portal username (email) for the Auvik connector. Leave blank to disable Auvik.
${user_config.auvik_username}cipp_client_idApp registration client ID for modern CIPP OAuth2 auth.
${user_config.cipp_client_id}cipp_tenant_idEntra tenant ID for modern CIPP OAuth2 auth.
${user_config.cipp_tenant_id}knowbe4_regionKnowBe4 shard: us (default), eu, ca, uk, de. Leave blank for us.
${user_config.knowbe4_region}vanta_base_urlOptional. Leave blank to use the default (https://api.vanta.com/v1).
${user_config.vanta_base_url}knowbe4_api_keyKnowBe4 reporting API key. Leave blank to disable KnowBe4.
${user_config.knowbe4_api_key}ninjaone_regionNinjaOne shard: us (default), eu, ca, oc. Leave blank for us.
${user_config.ninjaone_region}vanta_client_idVanta OAuth2 client ID. Leave blank to disable Vanta.
${user_config.vanta_client_id}blumira_base_urlOptional. Leave blank to use the default (https://api.blumira.com/public-api/v1).
${user_config.blumira_base_url}knowbe4_base_urlOptional. Leave blank to use the region default.
${user_config.knowbe4_base_url}spanning_api_urlOptional. Leave blank to use the platform default.
${user_config.spanning_api_url}blumira_client_idBlumira OAuth2 client ID. Leave blank if using a JWT token.
${user_config.blumira_client_id}blumira_jwt_tokenPre-issued Blumira JWT. Leave blank if using OAuth client credentials.
${user_config.blumira_jwt_token}ninjaone_base_urlOptional. Leave blank to use the region default; only set for a non-standard shard.
${user_config.ninjaone_base_url}paylocity_sandboxOptional. Set to 'true' to target the Paylocity sandbox (ignored if a base URL is set).
${user_config.paylocity_sandbox}spanning_platformSpanning platform: m365 (default) or google. Leave blank for m365.
${user_config.spanning_platform}cipp_client_secretApp registration client secret for modern CIPP OAuth2 auth.
${user_config.cipp_client_secret}cw_manage_base_urlOptional. Leave blank to use the default (https://api-na.myconnectwise.net).
${user_config.cw_manage_base_url}ninjaone_auth_modeOptional. Overrides the default OAuth flow only if your NinjaOne integration requires it.
${user_config.ninjaone_auth_mode}ninjaone_client_idNinjaOne OAuth2 client ID. Leave blank to disable NinjaOne.
${user_config.ninjaone_client_id}paylocity_base_urlOptional. Leave blank to use the default (https://api.paylocity.com).
${user_config.paylocity_base_url}spanning_api_tokenSpanning API token from the Spanning admin console.
${user_config.spanning_api_token}cw_manage_client_idConnectWise Manage integrator client ID from developer.connectwise.com.
${user_config.cw_manage_client_id}paylocity_client_idPaylocity OAuth2 client ID. Leave blank to disable Paylocity.
${user_config.paylocity_client_id}vanta_client_secretVanta OAuth2 client secret.
${user_config.vanta_client_secret}cw_manage_company_idConnectWise Manage company ID. Leave blank to disable ConnectWise.
${user_config.cw_manage_company_id}cw_manage_public_keyConnectWise Manage API member public key.
${user_config.cw_manage_public_key}paylocity_company_idPaylocity company ID that scopes all calls.
${user_config.paylocity_company_id}spanning_admin_emailSpanning admin email. Leave blank to disable Spanning.
${user_config.spanning_admin_email}threatlocker_api_keyThreatLocker API user key. Leave blank to disable ThreatLocker.
${user_config.threatlocker_api_key}blumira_client_secretBlumira OAuth2 client secret.
${user_config.blumira_client_secret}cw_manage_private_keyConnectWise Manage API member private key.
${user_config.cw_manage_private_key}threatlocker_base_urlOptional. Leave blank to use the default shard.
${user_config.threatlocker_base_url}ninjaone_client_secretNinjaOne OAuth2 client secret.
${user_config.ninjaone_client_secret}paylocity_client_secretPaylocity OAuth2 client secret.
${user_config.paylocity_client_secret}threatlocker_organization_idOptional. Only set when acting on a managed child organization.
${user_config.threatlocker_organization_id}Executes bash commands
Executes bash commands
Hook triggers when Bash tool is used
Hook triggers when Bash tool is used
Modifies files
Hook triggers on file write and edit operations
Modifies files
Hook triggers on file write and edit operations
Requires secrets
Needs API keys or credentials to function
Requires secrets
Needs API keys or credentials to function
Uses power tools
Uses Bash, Write, or Edit tools
Uses power tools
Uses Bash, Write, or Edit tools
Atlas is a self-configuring Claude Code plugin that turns any coding agent into a
disciplined multi-agent architect. Run /atlas once to onboard a project, then
drive work through the 20 auto-trigger skills and the atlas:<role> subagent
squad. A SessionStart hook loads the runtime every session, and four
self-improvement hooks (memory capture, auto-skill, nudge, session ingest) make
the agent better the more it is used in a codebase
(plugins/atlas/skills/atlas/SKILL.md:10,
plugins/atlas/README.md:3-9,
plugins/atlas/hooks/hooks.json:1-7).
The marketplace is published from this repo at
.claude-plugin/marketplace.json (Claude Code, name atlas, version 3.0.0,
.claude-plugin/marketplace.json:1-3) and at
.kimi-plugin/marketplace.json (Kimi Code CLI, version 2,
.kimi-plugin/marketplace.json:1-2). The Claude Code manifest currently ships
two plugins: atlas and the optional armada org-deployment plugin
(.claude-plugin/marketplace.json:18,28). The Kimi manifest ships the
same two plugins (atlas, armada) from .kimi-plugin/marketplace.json.
Two separate version counters, not a typo: the marketplace wrapper
above (3.0.0) versions the catalog file itself; the atlas plugin it
lists versions independently at 5.1.1
(plugins/atlas/.claude-plugin/plugin.json:3). Both numbers moved in the
same commit (ad7313c: marketplace 2.0.0 -> 3.0.0, plugin -> 5.0.0)
but on unrelated scales, so every v5.x reference later in this README
is the plugin version, never the marketplace version.
A separate armada plugin in this repo carries 11 department agents and 156
department skills for org deployment; install it alongside atlas only for org
use (plugins/atlas/README.md:8-10,
plugins/atlas/.claude-plugin/plugin.json:3).
.claude-plugin/marketplace.json via the /plugin command.atlas plugin from the marketplace (install armada too if
you want the department skills)./atlas once. The atlas-setup skill scaffolds
docs/ (plus .atlas/ internal state), verifies or installs claude-mem
and context-mode, wires hooks, and recommends the next step
(plugins/atlas/skills/atlas/SKILL.md:24-64,
plugins/atlas/skills/atlas-setup/SKILL.md:78-111).atlas-feature to build a feature, atlas-debug to root-cause a bug,
atlas-audit to run a code or architecture audit). The 20 non-setup skills
auto-trigger from the description field; you can also name them
directly (plugins/atlas/skills/atlas-setup/SKILL.md:150-162)..kimi-plugin/marketplace.json).scripts/ tooling
(plugins/atlas/hooks/hooks.json).claude-mem and context-mode are required companions. atlas-setup
detects them and offers to install if missing; do not install silently
(plugins/atlas/skills/atlas/SKILL.md:24-32).atlas itself. The
vendor MCP servers under mcp_servers/ are not part of the active
marketplace and require their own setup; see
.env.template if you use any of them.Top-level layout, one line each, every entry verified on disk:
Drive the Clippy macOS clipboard manager from Claude Code. Ships a self-contained MCP server (search, add, get, delete clips; manage categories), slash commands, clipboard-triage and clip-capture skills, and a clip-curator agent for organizing clipboard history.
Self-configuring multi-agent coding architect with 28 skills (11 Greek-mythology core skills: atlas-metis, atlas-hephaestus, atlas-ariadne, atlas-athena, atlas-argus, atlas-chronos, atlas-hermes, atlas-odysseus, atlas-nestor, atlas-armada, atlas-olympus; plus 17 task skills such as atlas-feature, atlas-debug, atlas-doctor, atlas-validate, atlas-refactor, atlas-harden, atlas-launch), 23 agents (12 core + 11 armada department), 15 hooks, and the atlas-metis method (evidence before any done claim). The atlas-armada skill provides organizational deployment with 11 departments, each with their own skills, agents, and vendor connectors. The atlas-olympus skill scaffolds the .atlas/ workspace, migrates the SSOT to .atlas/docs/, and recommends what to run next. All capabilities are now implemented as Claude Code skills following Anthropic's skill best practices.
Comprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.
Upstash Context7 MCP server for up-to-date documentation lookup. Pull version-specific documentation and code examples directly from source repositories into your LLM context.
Consult multiple AI coding agents (Gemini, OpenAI, Grok, Perplexity, plus codex, antigravity, and grok CLIs when installed) to get diverse perspectives on coding problems
Develop, test, build, and deploy Godot 4.x games with Claude Code. Includes GdUnit4 testing, web/desktop exports, CI/CD pipelines, and deployment to Vercel/GitHub Pages/itch.io.
A growing collection of Claude-compatible academic workflow bundles. Covers scientific figures, manuscript writing and polishing, reviewer assessment, citation retrieval, data availability, paper reading, literature search, response letters, paper-to-PPTX conversion, and evidence-grounded Chinese invention patent drafting. Rules are organized as reusable skill folders with explicit workflows and quality checks.
Harness-native ECC plugin for engineering teams - 67 agents, 279 skills, 94 legacy command shims, reusable hooks, rules, MCP conventions, and operator workflows for Claude Code plus adjacent agent harnesses