{
  "markdown": "# Maxim — the behavioral intelligence layer for Claude\n\n> The structural moat behind every Claude output. 91 specialist agents · 86 peer-reviewed frameworks (behavioral core + security · compliance · enterprise-architecture · engineering) · 14 compliance frameworks · 9 MCPs · 95 tools. Framework citation on every output. Drift caught before you ship. Voice locked across every surface.\n\n![Version](https://img.shields.io/badge/version-1.3.9.1-blue)\n![Agents](https://img.shields.io/badge/agents-91-green)\n![Frameworks](https://img.shields.io/badge/frameworks-86-orange)\n![Skills](https://img.shields.io/badge/skill_domains-52-purple)\n![Commands](https://img.shields.io/badge/commands-50-yellow)\n![MCP](https://img.shields.io/badge/MCP_servers-9-teal)\n![Tools](https://img.shields.io/badge/MCP_tools-95-teal)\n![Hooks](https://img.shields.io/badge/hooks-16-gray)\n![Drift_Classes](https://img.shields.io/badge/drift_classes-13-red)\n![License](https://img.shields.io/badge/license-BSL_1.1-lightgrey)\n\n---\n\n## Install · then see the moat work in 30 seconds\n\n```bash\n/plugin marketplace add DrNabeelKhan/maxim\n/plugin install maxim@maxim-packs\n```\n\nThen activate your tier (default: **90-day Trial of all 14 packs · no card · cancel anytime**):\n\n```bash\nbash bootstrap/install-tier-packs.sh        # Mac · Linux · WSL · Git Bash\npwsh -File bootstrap/install-tier-packs.ps1 # Windows · or PS7 cross-platform\n```\n\nThe wizard pre-selects Trial because **a moat is hard to evaluate when you can't see it**. Run your real work through Maxim for three months. Then decide what's worth keeping.\n\n> Per ADR-019 (Multi-Tenant Readiness): every operator starts with the same first-run experience. Pre-existing operators (like Maxim's maintainer running `nk-writer`) keep their advanced configurations untouched.\n\n---\n\n## What Maxim actually IS\n\nMost \"AI plugins\" give you templates. Some give you a chat wrapper. Maxim gives you a **governed multi-agent operating system** with structural enforcement of:\n\n1. **Framework citation on every output** — every Maxim emission names the behavioral science framework that justifies it (Fogg · Cialdini · Prospect Theory · OWASP · NIST · 70 more). The `behavioral-overlay-orchestrator` applies the ADR-007 citation discipline, and on the Claude Code CLI a commit-time gate rejects uncited pack output. The \"looks good\" generic LLM output is what Maxim is built to keep out.\n2. **Confidence tag on every output** — every emission carries 🟢 HIGH / 🟡 MEDIUM / 🔴 LOW per ADR-010's Technical Educator rubric. You see grounding depth, not just an answer.\n3. **Compliance overlay you don't have to remember** — 14 jurisdictional frameworks (GDPR · PIPEDA · UAE-PDPL · HIPAA · PCI-DSS · SOC2 · ISO 27001/13485/14971 · NIST CSF · EU AI Act · CASL · FINTRAC · WCAG 2.1) engage on regulated-data signals. The CSO auto-loop is a standing dispatch rule — reinforced by the default-on router on the Claude Code CLI — that even super-user mode keeps on for compliance work.\n4. **Voice routing as a property of agent invocation** — your voice loads when a writing agent dispatches, not when you remember to invoke `/mxm-brand-voice` first. Per ADR-016. Per operator via the v1.3.0 template pattern (ADR-019).\n5. **Proactive Watch drift detection** — 13 universal drift classes scan at session start on the Claude Code CLI, and on demand anywhere via `/mxm-watch`. Stale docs · broken refs · secret leaks · surface-claims drift · behavioral-moat drift. AI-coded projects rot fast; Watch is the rot detector.\n6. **Documents as Executable Contracts (ADR-002)** — CHANGELOG · MOAT_TRACKER · BUG_TRACKER · AGENT_SKILL_INVENTORY are read by the pre-commit hook as live state on the Claude Code CLI. Drift between claim and reality blocks the commit.\n\n**None of the external sources** (alirezarezvani · ui-ux-pro-max · superpowers · VoltAgent · planning-with-files) do any of this structurally. Maxim is the only layer that does. The community-pack system (ADR-008) composes Maxim's behavioral layer on top of their craft.\n\n---\n\n## See the moat work · 6 concrete use cases\n\n### Use Case 1 · Founder evaluating regulatory exposure\n\n```\nYou: \"We're adding stablecoin payments to our pricing flow. What jurisdictions\n      need DPAs and what's the controller/processor split?\"\n```\n\n**What fires:** `executive-router` → `cso-office` → `compliance-orchestrator` + `gdpr-counsel` + `hipaa-counsel` + `dpia-specialist`. MCPs invoked: `mxm-compliance.check_compliance` scans 14 frameworks · `mxm-compliance.generate_ropa_entry` stubs the ROPA · `mxm-compliance.get_jurisdiction_requirements` per region.\n\n**What you get:** jurisdictional map, cited per region, with DPA requirements, controller/processor split, and a flagged ROPA entry. 🟢 HIGH tag attached only if every jurisdiction in your `compliance.frameworks` returned PASS.\n\n**Without Maxim:** 2-3 days research, $5K+ in consultant fees, no audit trail.\n\n---\n\n### Use Case 2 · Engineer building a RAG pipeline with TDD discipline\n\n```\nYou: \"/mxm-build a RAG pipeline for our customer support docs\"\n```\n\n**What fires:** `cto-office` → `implementer` (lead) + `rag-specialist` (embodied via mxm-catalog) + `tester` + `reviewer`. MCPs invoked: `mxm-context.get_architecture_docs` surfaces existing ADRs · `mxm-behavioral.recommend_frameworks` applies TDD + BDD + C4 + arc42 · `mxm-behavioral.apply_framework Fogg-B-MAP` scope-checks before code touches.\n\n**What you get:** architecture with ADR-rated decisions, TDD scaffolding (tests first per Coverage Matrix), framework citations on every design choice. PII detected in the doc corpus auto-loops `cso-office.compliance-orchestrator`.\n\n**Without Maxim:** no framework citation, no scope check, security review post-PR (or never).\n\n---\n\n### Use Case 3 · PM writing a PRD with Fogg B=MAP + JTBD\n\n```\nYou: \"/mxm-pm prd new operator onboarding flow\"\n```\n\n**What fires:** `cpo-office` → `product-manager` + `onboarding-designer`. MCPs invoked: `mxm-behavioral.apply_framework Fogg-B-MAP` returns motivation/ability/prompt analysis · `mxm-behavioral.apply_framework JTBD` returns job statements via Ulwick's Jobs Atlas · `mxm-catalog.get_handoff_chain` returns cross-office collaboration map for the rollout.\n\n**What you get:** PRD with Fogg B=MAP scope check, JTBD job statements (functional + emotional + social), RICE prioritization, and explicit collaboration handoffs to CMO (announcement) + CTO (implementation) + CSO (data-handling).\n\n**Without Maxim:** PRD template with no framework grounding, scope creeps mid-sprint, RICE done by gut.\n\n---\n\n### Use Case 4 · Marketer creating a research-backed launch podcast in operator voice\n\n```\nYou: \"Create a podcast about our v1.3 launch using these customer interviews\n      and product docs\" + [drop URLs/PDFs]\n```\n\n**What fires:** `executive-router` → `cmo-office` → `notebooklm-content-production` (ADR-018 routing) → `{your-operator-id}-writer` for the intro/outro. MCPs invoked: `mxm-notebooklm.notebook_create` → `source_add_url` × N → `source_wait` → `generate_audio_overview format=deep-dive` → `artifact_wait` → `artifact_download`. Operator voice loaded via `.brand-foundation/personal.local/` (ADR-019 template pattern).\n\n**What you get:** 20-min podcast in YOUR voice, with intro/outro in operator-voice (the same voice as your blog posts), citations from your source material, audit trail showing every framework that fired (Diátaxis explanation mode · Diffusion of Innovations multi-format generation).\n\n**Without Maxim:** hand-write script, generic TTS, voice drift across artifacts.\n\n---\n\n### Use Case 5 · Session-end ritual you forgot you need\n\n```\nYou: \"/mxm-session-end\"\n```\n\n**What fires:** `coo-office` → `planner` orchestrates the 9-document closure bundle. MCPs invoked: `mxm-portfolio.sync_portfolio` syncs project metrics · `mxm-memory.archive_session_memory` persists session to MemPalace · `mxm-context.watch_run` runs LIGHT drift scan across 13 classes.\n\n**What you get:** SESSION_CONTINUITY.md updated · session-YYYY-MM-DD.md appended · CHANGELOG entry if user-facing change · MOAT_TRACKER row if positioning changed · AGENT_SKILL_INVENTORY refreshed if capabilities touched · project-manifest `last_activity` bumped · skill-gaps log appended. Per CLAUDE.md \"Session without memory writes = session wasted.\"\n\n**Without Maxim:** session memory lost, next session starts cold, drift accumulates.\n\n---\n\n### Use Case 6 · Auto-loop you didn't ask for (and saved you from a breach)\n\n```\nYou: \"Add user authentication to our app\"\n```\n\n**What fires:** `cto-office` → `backend-architect` (primary) **automatically loops** → `cso-office` → `appsec-engineer` + `secure-code-reviewer` + `owasp-specialist`. MCPs invoked: `mxm-compliance.check_compliance` fires on the regulated-data signal (auth = PII) · `ethics-orchestrator` validates · OWASP Top 10 + LLM Top 10 frameworks cited inline.\n\n**What you get:** auth implementation AND a parallel security review in the same conversation. Framework citations (OWASP Top 10 § A07:2021 Identification and Authentication Failures · NIST CSF PR.AC). 🔴 LOW tag and BLOCK if a known anti-pattern detected (plain-text password storage, missing rate-limiting, JWT without expiry).\n\n**Without Maxim:** security review happens at PR time (best case) or never (typical case).\n\n---\n\n## The 9 MCPs at a glance · 95 tools you can call directly\n\n| MCP | Tools | What it does |\n|---|---:|---|\n| **mxm-portfolio** | 9 | Project state, sync across 21+ projects, portfolio metrics |\n| **mxm-context** | 15 | Architecture docs · design refs · 13-class drift detection |\n| **mxm-catalog** | 9 | Agent + office + skill + command catalog, L2 specialist descent (route_task) |\n| **mxm-compliance** | 5 | 14 jurisdictional frameworks · ROPA entries · per-jurisdiction requirements |\n| **mxm-behavioral** | 7 | 86 behavioral frameworks · recommend · apply · score moat coverage |\n| **mxm-memory** | 6 | Session history · decision log · cross-project memory |\n| **mxm-voice** | 4 | Voice-driven office routing (wraps mbailey/voicemode for STT+TTS) |\n| **mxm-commands** | 2 | All 50 slash commands as MCP tools (cross-surface command parity) |\n| **mxm-notebooklm** | 38 | NotebookLM research synthesis (wraps teng-lin/notebooklm-py MIT · v1.2.1.0+) |\n\n> v1.3.2 surface-claims-drift correction: prior README declared 87 tools by undercounting 4 MCPs (behavioral · context · memory · portfolio) by 2 each. Caught by pre-release-audit ground-truth grep against `server.tool(` declarations. Source-of-truth: `documents/ledgers/AGENT_SKILL_INVENTORY.md`.\n\n**Cross-surface:** all 9 MCPs work on Claude Code · Desktop · Cowork. Claude.ai Web reaches them via MCP-over-API.\n\n---\n\n## The trial · why we default to it\n\nYou can't evaluate a moat you can't see. The 90-day trial unlocks all 6 L1 packs (the structural moats) plus L2 vertical bundles plus L3 industry packs. Run your real work through Maxim for three months:\n\n- **Week 1-2:** install runs, you trigger your first auto-loops\n- **Week 3-5:** drift detection catches something you didn't know was broken (Class 11 surface-claims-drift is the typical first surprise)\n- **Week 6-9:** behavioral framework citations start changing your decisions — you see WHICH framework Maxim applied to which output, you understand why\n- **Week 10-12:** you have data — which packs you actually used, which fired auto-loops on your real work, which compliance gates blocked work that would have been broken\n\nBy day 90 you know exactly which tier matches your work. Convert or downgrade without guessing.\n\n---\n\n## Tier roadmap\n\n| Tier | Includes | Best for |\n|---|---|---|\n| **Solo** | Core (free forever) | Solo operators starting out · evaluating Maxim |\n| **Pro** | Core + 6 L1 packs | Serious operators on 1-2 projects |\n| **Team** | Core + L1 + 4 L2 verticals | Teams across founder · growth · pro · agency |\n| **Enterprise** | All 14 packs (incl. L3 healthcare · legal · fintech · govtech) | Regulated industries · multi-team orgs |\n| **Trial** | All 14 packs · 90 days · no card | Anyone evaluating |\n\nPricing at [maxim.isystematic.com/pricing](https://maxim.isystematic.com/pricing) — you decide after the trial, not before.\n\n---\n\n## For specific roles · use the persona dispatchers\n\nMaxim ships 5 TIER 3 persona dispatchers that speak your vocabulary:\n\n```\n/mxm-legal     jurisdictional-map · privacy-impact · contract-review · vendor-dpa · regulatory-map\n/mxm-arch      capability-map · wardley-map · tech-radar · c4-diagram · adr · vendor-eval\n/mxm-secure    threat-model · owasp · sbom · incident · compliance-posture · ai-risk\n/mxm-founder   pitch-deck · gtm-plan · runway-model · pricing · business-model-canvas · competitive-moat\n/mxm-pm        prd · user-story · okr · prioritize · jtbd\n```\n\nEach persona routes to the right specialist within Maxim's 91-agent roster automatically. Legal pros think \"DPIA,\" not \"CSO compliance skill.\" Architects think \"Wardley map,\" not \"CEO enterprise-architect office.\" TIER 3 commands speak the persona's language and route invisibly.\n\n---\n\n## Multi-surface deployment\n\nMaxim runs everywhere Claude does:\n\n| Surface | Fidelity | What works |\n|---|---|---|\n| **Claude Code** | 100% | All 50 commands · 9 MCPs (95 tools) · 24 dispatchable subagents · all behavioral overlays |\n| **Claude Desktop** | ~95% | 9 MCPs (95 tools) · paste `maxim-project-instructions.md` for behavioral layer |\n| **Claude.ai Web** | ~85% | Project instructions · MCP-over-API when available |\n| **Claude.ai Cowork** | ~85% | Plugin bundles MCPs natively |\n\nDesktop one-command setup: `bash bootstrap/mxm-desktop-config.sh` or `pwsh -File bootstrap/mxm-desktop-config.ps1`.\n\n---\n\n## Architecture decisions · 22 ADRs (18 public · 4 confidential)\n\nMaxim ratifies every architectural choice via ADRs. The 18 public ones describe Maxim's commitments to operators:\n\n- [ADR-002](documents/ADRs/ADR-002-documents-as-executable-contracts.md) Documents as Executable Contracts (the structural rule)\n- [ADR-004](documents/ADRs/ADR-004-free-tier-executable-contract.md) Free tier specification\n- [ADR-007](documents/ADRs/ADR-007-behavioral-moat-framing-doctrine.md) Behavioral Moat Framing Doctrine\n- [ADR-008](documents/ADRs/ADR-008-community-pack-system.md) Community Pack System\n- [ADR-009](documents/ADRs/ADR-009-pack-architecture-l1-l2-l3.md) Pack Architecture (6 L1 + 4 L2 + 4 L3)\n- [ADR-010](documents/ADRs/ADR-010-confidence-tag-technical-educator-rubric.md) Confidence Tag rubric\n- [ADR-016](documents/ADRs/ADR-016-voice-writing-agent-architecture.md) Voice Writing Agent Architecture\n- [ADR-017](documents/ADRs/ADR-017-office-as-dispatch-boundary.md) Office-as-Dispatch-Boundary\n- [ADR-018](documents/ADRs/ADR-018-external-tool-integration-pattern.md) External Tool Integration Pattern\n- [ADR-019](documents/ADRs/ADR-019-multi-tenant-readiness.md) Multi-Tenant Readiness (v1.3.0)\n- [ADR-021](documents/ADRs/ADR-021-maxim-default-on-router.md) **Maxim Default-On — always-on intent router (v1.3.5)**\n- plus ADR-011 · ADR-012 · ADR-013 · ADR-014 · ADR-015 (see [INDEX.md](documents/ADRs/INDEX.md))\n\nFour additional ADRs cover internal architecture (dispatch baseline · IP protection · Worker license issuance · external content boundary) and remain in the operator's private ledger.\n\n---\n\n## License · community · support\n\n**License:** BSL-1.1 — converts to Apache 2.0 after 4 years per ADR-005. Permissive enough to build on, structured enough to fund development.\n\n**Issues:** [github.com/DrNabeelKhan/maxim/issues](https://github.com/DrNabeelKhan/maxim/issues)\n**Pricing:** [maxim.isystematic.com/pricing](https://maxim.isystematic.com/pricing) (decide after the trial)\n**Docs:** [HELP.md](documents/guides/HELP.md) for the full command catalog · [GETTING_STARTED.md](documents/guides/GETTING_STARTED.md) for onboarding · [INSTALL.md](documents/INSTALL.md) for multi-surface deployment\n\n---\n_Copyright (c) 2026 iSystematic Inc. Maxim is a product of iSystematic Inc. Licensed under Business Source License 1.1._",
  "bytes": 15960,
  "sha": "39cf42072caf5c3e942923a5b279fb2d9955e1442e53c01423ae441d0c029bad",
  "repo_slug": "drnabeelkhan/maxim",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/plg_drnabeelkhan_maxim_maxim_383ecd7b/readme"
}