Claude Code 32-Agent Team

A production-grade multi-agent engineering team for Claude Code, with hook-enforced protocol, adversarial review, Bayesian trust calibration, dynamic specialist hiring, an optional parallel cognitive layer, and a self-improving meta-cognitive loop.

Architecture diagrams: ARCHITECTURE_DIAGRAMS.md (editable Mermaid source + ASCII fallbacks + PNG export instructions)

This is not a framework. It's an operating system layer on top of Claude Code — 32 domain-expert agents coordinated through a syscall protocol (NEXUS), governed by runtime hooks, validated by a 352-assertion contract test suite, calibrated by a per-agent trust ledger, and equipped with a gated hiring pipeline for growing the roster as coverage gaps appear.

What it actually is (verified against itself):

• 32 agent definitions in agents/ (each 20–56 KB of role-specific instructions, spanning Go / Python / TypeScript / Elixir / BEAM / K8s / GCP / security / QA / AI platform / external BEAM retainer / recruiting + team-evolution roles)
• Runtime hooks that enforce closing-protocol invariants, persist signals, auto-log syscalls, auto-record trust verdicts, and gate post-hire registration
• A signal bus (agent-memory/signal-bus/) that carries memory/evolution/cross-agent flags between sessions
• A trust ledger (agent-memory/trust-ledger/) that Bayesian-blends per-agent accuracy from validator verdicts
• A documented NEXUS syscall protocol for privileged operations (SPAWN, SCALE, RELOAD, MCP, ASK, CRON, INTUIT, etc.)
• A Dynamic Domain Expert Acquisition pipeline (talent-scout + recruiter) — the team detects its own coverage gaps (5-signal confidence score + session-sentinel co-sign) and hires new specialist agents through 8 gated phases (research → synthesis → validation → challenger review → atomic registration → post-hire verify → probation → promotion)
• An optional Shadow Mind cognitive layer (intuition-oracle + Observer + Pattern Computer + Speculator + Dreamer + Pattern Library) that offers probabilistic pattern-based guidance via [NEXUS:INTUIT], with topic cluster pattern retrieval ("have we seen this before? what fixed it?"), delete-to-disable without affecting the conscious team
• Execution modes (FAST / BALANCED / FULL_POWER) for speed/cost/quality trade-offs per task
• 6-state agent lifecycle (candidate → probationary → active → trusted → deprecated → retired) with trust-weighted routing preference
• Quality gates by task type — different workflows for code changes vs security findings vs architecture decisions vs incidents
• Conflict arbitration protocol — structured evidence-based dispute resolution when agents disagree
• nexus-doctor.sh — diagnostic health check for framework integrity (agents, hooks, dependencies, Shadow Mind status)
• Planning cache (playbooks) — reusable task workflows that make repeated task types faster over time
• Contract tests (tests/agents/) that run on every commit and gate prompt evolution

---

Why this exists

Most agentic frameworks try to enforce discipline inside the agent loop (via code wrapping the model). This team enforces it below the agents (via hooks at the Claude Code runtime layer). That's structurally better: agents cannot talk their way past a SubagentStop hook the way they can talk their way past a Python try/except.

The team's value proposition splits cleanly:

• Methodology layer (NEXUS + signal bus + validator gate + challenger gate + trust ledger + Pattern F + hiring pipeline + Shadow Mind): domain-agnostic, transfers to any engineering work
• Expertise layer (32 agent prompts covering Go / Python / TypeScript / Elixir / BEAM / K8s / GCP / security / QA / AI platform / etc.): extend or customize for your stack — or let the hiring pipeline add specialists as gaps appear

---

Installation

# 1. Copy the team into your project
cp -R claude-nexus-hyper-agent-team/agents           YOUR_PROJECT/.claude/agents/
cp -R claude-nexus-hyper-agent-team/hooks            YOUR_PROJECT/.claude/hooks/
cp -R claude-nexus-hyper-agent-team/tests            YOUR_PROJECT/.claude/tests/
cp -R claude-nexus-hyper-agent-team/docs             YOUR_PROJECT/.claude/docs/
cp -R claude-nexus-hyper-agent-team/agent-memory     YOUR_PROJECT/.claude/agent-memory/
cp    claude-nexus-hyper-agent-team/settings.json    YOUR_PROJECT/.claude/settings.json
cp    claude-nexus-hyper-agent-team/CLAUDE.md        YOUR_PROJECT/CLAUDE.md

# 2. Verify installation
python3 YOUR_PROJECT/.claude/tests/agents/run_contract_tests.py
# Expected: 352 passed, 0 failed

# 3. Optional — install the git pre-commit hook for agent contract tests
ln -s ../../.claude/hooks/pre-commit-agent-contracts.sh YOUR_PROJECT/.git/hooks/pre-commit

After install

1. Open CLAUDE.md at your project root and fill in the ## Project-Specific Context section at the bottom (architecture overview, development commands, environment variables, deployment notes).
2. Replace the `` tokens with your project's specifics:
   • <your project> → your project name
   • <your-project> → your project slug
   • <go-service> → your Go service name (or remove if not applicable)
   • `` → your Python service name (or remove if not applicable)
   • <frontend> → your frontend package name (or remove if not applicable)
   • `` → a filesystem-safe slug for your project
3. Start your first session — say "use the team, <goal>" and Claude Code will dispatch the CTO.

Customizing the roster

The 32 agents cover: Go, Python, TypeScript, React/Next.js, Elixir, BEAM/OTP, Kubernetes, GCP, Terraform, PostgreSQL, Redis, Firestore, GraphQL, security, observability, testing, AI/ML architecture, benchmarking, and external BEAM retainer advisory. If you work on a different stack (e.g., Rust, embedded, kernel, ML training, AWS-native), you have two paths:

**Path A — Let the team hire for you (r