"LoopBack" matching MCP servers:

Enables MCP clients to search, read, and record reviewed conclusions from policy-gated allowlisted files, with sanitized chunks and create-only drafts or append-only lessons, without granting general filesystem or execution access. It supports lexical ranking and optional loopback embeddings while rejecting likely credentials and raw challenge flags.

Enables Codex to retrieve read-only persona snapshots, relationship projections, and FTS5 public knowledge from a local loopback service, with privacy-safe bundle import and a management UI.

Provides persistent, local-first memory for coding agents with Markdown as the source of truth, exposed via CLI, loopback API, MCP, and Codex hooks for context retrieval and durable writes.

Enables routing of natural-language tasks to the best-matching skill in a local pool of SKILL.md files, using a two-stage bi-encoder retrieval plus pointwise reranking pipeline that runs entirely on your machine. Exposes tools for searching skills, fetching a full skill record, triggering incremental or forced reindexing, and inspecting index statistics over a loopback-hosted MCP endpoint.

Enables agents to store documents and atomic facts with provenance in a local SQLite knowledge base, then search them through fused keyword, exact, semantic, entity, graph and fact retrieval arms that return one ranked list where every result carries its address, trust level and a reason for its rank. A read-only terminal and loopback web view let humans inspect the same searches, timelines, entity neighbourhoods and agent-written pages without being able to alter the knowledge base.