Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Springdrift: An Auditable Persistent Runtime for LLM Agents with Case-Based Memory, Normative Safety, and Ambient Self-Perception

Seamus Brady

Abstract

We present Springdrift, a persistent runtime for long-lived LLM agents. The system integrates an auditable execution substrate (append-only memory, supervised processes, git-backed recovery), a case-based reasoning memory layer with hybrid retrieval (evaluated against a dense cosine baseline), a deterministic normative calculus for safety gating with auditable axiom trails, and continuous ambient self-perception via a structured self-state representation (the sensorium) injected each cycle without tool calls. These properties support behaviours difficult to achieve in session-bounded systems: cross-session task continuity, cross-channel context maintenance, end-to-end forensic reconstruction of decisions, and self-diagnostic behaviour. We report on a single-instance deployment over 23 days (19 operating days), during which the agent diagnosed its own infrastructure bugs, classified failure modes, identified an architectural vulnerability, and maintained context across email and web channels -- without explicit instruction. We introduce the term Artificial Retainer for this category: a non-human system with persistent memory, defined authority, domain-specific autonomy, and forensic accountability in an ongoing relationship with a specific principal -- distinguished from software assistants and autonomous agents, drawing on professional retainer relationships and the bounded autonomy of trained working animals. This is a technical report on a systems design and deployment case study, not a benchmark-driven evaluation. Evidence is from a single instance with a single operator, presented as illustration of what these architectural properties can support in practice. Implemented in approximately Gleam on Erlang/OTP. Code, artefacts, and redacted operational logs will be available at https://github.com/seamus-brady/springdrift upon publication.

Springdrift: An Auditable Persistent Runtime for LLM Agents with Case-Based Memory, Normative Safety, and Ambient Self-Perception

Abstract

We present Springdrift, a persistent runtime for long-lived LLM agents. The system integrates an auditable execution substrate (append-only memory, supervised processes, git-backed recovery), a case-based reasoning memory layer with hybrid retrieval (evaluated against a dense cosine baseline), a deterministic normative calculus for safety gating with auditable axiom trails, and continuous ambient self-perception via a structured self-state representation (the sensorium) injected each cycle without tool calls. These properties support behaviours difficult to achieve in session-bounded systems: cross-session task continuity, cross-channel context maintenance, end-to-end forensic reconstruction of decisions, and self-diagnostic behaviour. We report on a single-instance deployment over 23 days (19 operating days), during which the agent diagnosed its own infrastructure bugs, classified failure modes, identified an architectural vulnerability, and maintained context across email and web channels -- without explicit instruction. We introduce the term Artificial Retainer for this category: a non-human system with persistent memory, defined authority, domain-specific autonomy, and forensic accountability in an ongoing relationship with a specific principal -- distinguished from software assistants and autonomous agents, drawing on professional retainer relationships and the bounded autonomy of trained working animals. This is a technical report on a systems design and deployment case study, not a benchmark-driven evaluation. Evidence is from a single instance with a single operator, presented as illustration of what these architectural properties can support in practice. Implemented in approximately Gleam on Erlang/OTP. Code, artefacts, and redacted operational logs will be available at https://github.com/seamus-brady/springdrift upon publication.

Paper Structure

This paper contains 99 sections, 2 equations, 6 figures, 10 tables.

Table of Contents

  1. Introduction
  2. Contributions
  3. Scope
  4. Runtime Architecture
  5. Overview
  6. Configuration and Continuity
  7. Memory Subsystem
  8. Cognitive Cycle
  9. Specialist Agents
  10. Auditability and Git Backup
  11. Implementation
  12. The Sensorium: Continuous Self-Perception
  13. Motivation
  14. Design
  15. Performance Signals
  16. ...and 84 more sections

Figures (6)

  • Figure 1: System architecture. The cognitive loop orchestrates all components via typed message-passing channels. This is a simplified diagram; the system can also create teams of cooperating subagents.
  • Figure 2: Cognitive cycle. Every input passes through safety gates before and after inference.
  • Figure 3: Six-signal retrieval pipeline.
  • Figure 4: Learning curve. P@4 improves with case base size; hard queries benefit most (+17.7%).
  • Figure 5: D$'$ scorer to normative calculus pipeline.
  • ...and 1 more figures