field reference
behavioral_state_decay(t) — observed loss of instruction adherence, calibration, and consistency as agent sessions grow long.
Documenting, measuring, and mitigating behavioral degradation in long-running AI agents — definitions, metrics, and mitigation patterns from real operations.
Field Notes
Compacting a long-running agent's context to control token cost has a measured side effect: the agent starts re-fetching, re-checking, and re-exploring facts it already established before the compaction happened. What the 2026 evidence shows about why this happens, how to spot it in a trajectory log before it shows up as a bigger invoice, and how to bound it.
Definitions
Memory poisoning is the injection of false or adversarial content into an agent's persistent memory store, where it survives across sessions and is later retrieved as if it were verified fact. How it differs from context rot and drift, and why aggressive memory write policies make it worse.
Definitions
Calibration drift is the failure mode where an agent's expressed confidence in its own output separates from that output's actual correctness over the course of a session, growing more confident, less confident, or simply noisier without any change in how right the agent actually is. Distinct from sycophantic drift and context rot, grounded in multi-turn calibration research spanning self-anchoring, reasoning-budget effects, and mechanistic circuit analysis, with detection and mitigation patterns.
Definitions
Context handoff loss is the failure mode where a constraint, fact, or instruction established by one agent fails to survive the boundary into the next agent's context, silently producing a downstream action built on an incomplete picture. Distinct from coordination drift and context rot, grounded in a 150-trace human-annotated multi-agent failure taxonomy, with detection and mitigation patterns.
Measurement
Teams use an LLM judge to score drift regression suites at scale, but the judge carries its own unreliability and can silently change out from under a monitoring pipeline. What the 2026 evaluation literature measures, and how to tell a real system regression from a judge that moved.
Definitions
Tool-use drift is the failure mode where an agent's tool selection degrades over a session: it reaches for a recently-used or memory-favored tool instead of the one the current step actually needs, even when the correct tool is sitting in its own retrieved context. Grounded in 2026 benchmarks isolating selection failure from retrieval failure, with detection and mitigation patterns distinct from goal drift and role drift.