Aegis-Sovereign MCP Server
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
No arguments | |||
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| sovereign_optimize_contextA | Pre-filters large archives down to verified evidence chunks with citations. Cuts prompt tokens by 90%+ (guaranteed >= 40%) in a single call, eliminating cloud token waste. |
| sovereign_search_vaultB | Direct hybrid dense + sparse vector search across the sovereign vault. |
| sovereign_get_entity_dossierA | Compiles a complete relational knowledge graph dossier for an entity: linked documents, timeline of dates, monetary transactions, and multi-hop cross-references (1 to 5 hops). |
| sovereign_node_statusA | Inspects health, indexing stats, and operational metrics of the Sovereign Appliance node. |
| sovereign_route_and_analyzeC | Executes the Two-Pronged Hybrid Retrieval & Resilient Intent Router (ADR-40). Routes deterministic identifiers (<2ms SQLite B-Tree/FTS5 with MAC pushdown) vs multi-tier cognitive queries (hybrid_needle, relational_graph, macro_synthesis, compound_fused) with local NanoRunner synthesis and explicit execution_mode telemetry. |
| sovereign_inspect_archiveB | Streams and inspects .hdx, .hwics, .zip, .xlsx, .docx, .tar.zst, or .epub containers purely in-memory (O_RDONLY, zero disk extraction) per ADR-07. Also supports section/offset slicing (section_filter, char_offset) and zero-reindex corpus path relocation / stale relationship cleanup (action='relocate_prefix' | 'purge_prefix' | 'corpus_stats'). |
| sovereign_scan_onboarding_radarA | Runs the ADR-08 60-Second Auto-Discovery Onboarding Radar (os.scandir, strictly read-only) to discover, classify, and rank domain knowledge directories (fiscal_nfe, medical_clinical, legal_contracts, engineering_manuals, general_knowledge). |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 7 tools
Most tools target distinct operations (scan, status, optimize, search, dossier, route, inspect), but sovereign_optimize_context, sovereign_search_vault, and sovereign_route_and_analyze all involve retrieval/analysis and could be confused for similar high-level queries. The descriptions help, but the boundaries between 'optimize context', 'search vault', and 'route and analyze' are not immediately crisp.
All tools share the 'sovereign_' prefix and use verb_noun or verb_phrase patterns (scan_onboarding_radar, node_status, optimize_context, search_vault, get_entity_dossier, route_and_analyze, inspect_archive). Minor inconsistency: 'sovereign_node_status' lacks a verb, while others have explicit verbs, but overall the pattern is predictable.
Seven tools is a reasonable count for a specialized knowledge-management/retrieval server. The scope is broad (onboarding, search, routing, archive inspection, entity dossiers), but each tool covers a distinct capability without feeling bloated.
The server covers discovery, search, retrieval, routing, archive inspection, and node health, but lacks obvious lifecycle operations like adding/updating/deleting documents or managing the vault corpus. The inspect_archive tool has some maintenance actions (relocate_prefix, purge_prefix), but there is no clear ingest/write tool, leaving a notable gap for a knowledge vault server.