Seven architectural domains. One unified platform. From plugin ecosystems to TRiSM-compliant governance — everything you need to deploy production-grade AI agent infrastructure.
Each domain is a self-contained architecture with detailed requirements, design decisions, and implementation blueprints.
Extensible plugin architecture with 8 lifecycle hook events, slash commands, skills framework, agent definitions, and MCP server integration. Transform Claude Code from a CLI into a programmable platform.
5-signal tier classification routes tasks through specialized agents with quality gates. TRIVIAL to MAJOR workflows with BRD-driven development, intent engineering, and adversarial dual-AI review.
Hierarchical CLAUDE.md management with auto-memory, session persistence, and intelligent context compression. Your agents maintain coherent behavior across sessions and projects.
Cross-session knowledge with Qdrant vector storage, semantic recall, and automated lifecycle management. Store, recall, organize, and forget — agents that learn from every interaction.
Trust levels, data classification, audit trails, LLM threat detection, and SIEM integration. 15 event types, manifest-based identity, and approval gates for enterprise-grade AI deployment.
Interactive web UI with D3.js knowledge graphs, Chart.js analytics, semantic search, and collection monitoring. See your memory system in real-time with drift detection and relationship visualization.
HTML-to-markdown proxy that gives agents clean, token-counted content from any URL. Trafilatura conversion with Playwright fallback for JS-heavy sites and graceful degradation pipeline.
Not templates. Not downloads. Architectural blueprints that generate implementations tailored to your environment.
Every production environment has unique constraints — server topology, security policies, compliance requirements, workflow patterns. A single implementation can't fit all of them.
Platinum Goose provides the architecture. You provide the environment. Claude Code generates the implementation.
// How it works:
1. Read the PRD for your domain
2. Identify which requirements apply
3. Copy the Claude Code prompt
4. Generate tailored implementation
5. Deploy to your environment
// Example: Add governance to existing system
$ claude
> Build a governance framework using the
Platinum Goose PRD. My environment:
- 12 custom agents
- Qdrant on Docker (port 6333)
- n8n for automation
- Wazuh for SIEM
// Claude generates your implementation:
manifest.py ← agent identity
trust_broker.py ← delegation rules
policy_engine.py← tool gating
audit_bus.py ← event store
alerting.py ← Wazuh integrationSeven comprehensive PRDs. Hundreds of requirements. Complete architectural blueprints for production AI systems.