Wikantik Platform Development Hub: Architecture, MCP Protocols, and Agent Briefings

Wikantik is a high-performance knowledge platform and Agentic Memory engine designed to bridge human technical authoring with autonomous AI coding agents. Built on a modernized Java/JSPWiki core with Lucene hybrid search, Jena SPARQL RDF Knowledge Graphs, and Model Context Protocol (MCP) tool surfaces, Wikantik serves as both an enterprise wiki and a real-time retrieval engine for AI agent sessions.

This hub details Wikantik platform internals, MCP integration patterns, search pipeline architectures, and developer extension guides.

---

1. Quick-Reference: Wikantik Architecture Stack

+-----------------------------------------------------------------------------------------+
|                               WIKANTIK PLATFORM ARCHITECTURE                            |
+-----------------------------------------------------------------------------------------+
| Layer                  | Technologies                      | Primary Functionality      |
+------------------------+-----------------------------------+----------------------------+
| Agent Surface          | MCP Server (`/wikantik-admin-mcp`)| 21+ autonomous tool actions|
| Knowledge Graph        | Apache Jena RDF / SPARQL          | Semantic triple store      |
| Hybrid Search Engine   | Apache Lucene + Vector Embeddings | Keyword & vector retrieval |
| Storage & Versioning   | Git-backed Markdown Repository    | Change history, blame, WAL |
| Web UI & REST API      | Java 21 / React Micro-frontends   | Collaborative curation     |
+-----------------------------------------------------------------------------------------+

---

2. Core Subsystems & Developer Guides

I. Agent Integration & MCP Tools

• WikantikAgentContextGuide: Comprehensive guide on session briefings (assemble_bundle, get_briefing) and the 4-tier MCP escalation ladder.
• McpAuditTools: Automated diagnostics and regression testing for MCP tool registrations.
• FindingTheRightMcpTool: Decision tree for selecting retrieval, authoring, and graph MCP tools.

II. Knowledge Graph & Semantic Metadata

• KnowledgeGraphDogfooding: Using RDF triples to navigate internal wiki relationships and ontology consistency.
• GraphProjector: Transforming markdown frontmatter links into Jena RDF model triples.
• FrontmatterConventions: Standardized schema validators for cluster, summary, and audience metadata.

III. Search & Retrieval Quality

• EvaluatingRetrievalQuality: Precision@K, Recall@K, and MRR benchmarking for hybrid wiki search.
• InterpretingHybridRetrievalMetrics: Diagnosing reciprocal rank fusion (RRF) scores across BM25 and vector embeddings.