# Harness engineering glossary > The vocabulary of the agent harness — the prompts, tools, skills, evals, and memory that wrap the model call — and of the work of improving it from production traces. Canonical: https://moda.dev/glossary Updated: 2026-08-29 When an agent misbehaves in production, the fix almost always lands in the harness layer, not the model weights. These are the terms teams use while doing that work. Written by the team building Moda, a harness engineering platform that turns production traces into verified improvements for the agent harness. ## Agent harness Everything that wraps the model call to produce agent behavior: system prompts, tool definitions and routing, workflow orchestration, retrieval, memory, evals, and guardrails. Glossary entry: https://moda.dev/glossary#agent-harness ## Agent laziness A behavioral failure where the agent declines, hedges, or gives a generic answer to a task it is capable of completing. Glossary entry: https://moda.dev/glossary#agent-laziness ## AI agent observability In a harness engineering workflow, the input layer: the conversation-level record of what production agents did. The record only pays off when it becomes a specific harness change. Glossary entry: https://moda.dev/glossary#agent-observability ## Behavioral failure A failure mode that occurs at the conversation level, not at the API call level — invisible to traces alone. Glossary entry: https://moda.dev/glossary#behavioral-failure ## Cluster hierarchy Three-level segment clustering — Category → Subcategory → Cluster — built with HDBSCAN over Qwen3 embeddings. Glossary entry: https://moda.dev/glossary#cluster-hierarchy ## Continual learning Harness-layer improvement from production signal: updating prompts, tools, workflows, retrieval, memory, and evals so the learnings live outside the model weights and apply across whichever model the harness mounts. Glossary entry: https://moda.dev/glossary#continuous-learning ## Context loss A behavioral failure where the agent forgets earlier turns, contradicts itself, or asks the user to re-supply known information. Glossary entry: https://moda.dev/glossary#context-loss ## Conversation analytics Analytics that treat full agent conversations — not individual calls — as the unit of analysis. Glossary entry: https://moda.dev/glossary#conversation-analytics ## Frustration root cause For every detected frustration event, the trigger turn, leading trajectory, affected user goal, and counterfactual for what the agent should have done. Glossary entry: https://moda.dev/glossary#frustration-root-cause ## Goal drift A behavioral failure where the agent silently abandons the user's original objective and solves a different problem. Glossary entry: https://moda.dev/glossary#goal-drift ## Hallucination When the agent fabricates facts, invents tool calls, or claims to have performed actions it did not perform. Glossary entry: https://moda.dev/glossary#hallucination ## Intent discovery Hierarchical clustering of production conversations into a 3-level taxonomy with no manual tagging. Glossary entry: https://moda.dev/glossary#intent-discovery ## LLM observability Call- and trace-level monitoring of LLM applications — prompts, completions, tokens, latency, errors. It sits upstream of the harness layer: it records what happened; harness engineering decides what to change. Glossary entry: https://moda.dev/glossary#llm-observability ## LLM tracing Recording the step-by-step execution of an LLM application — every prompt, tool call, and response — for debugging. Glossary entry: https://moda.dev/glossary#llm-tracing ## MCP server A server that exposes tools and resources to AI coding assistants via the Model Context Protocol. Glossary entry: https://moda.dev/glossary#mcp-server ## Moda Moda is a harness engineering platform for teams with an AI agent already in production. It turns production traces into verified improvements for the agent harness: prompts, tools, skills, evals, memory. Glossary entry: https://moda.dev/glossary#moda ## OpenLLMetry An open-source SDK that emits OpenTelemetry-compatible traces for LLM and agent applications. Glossary entry: https://moda.dev/glossary#openllmetry ## OpenTelemetry The industry-standard open-source framework for emitting and collecting telemetry — traces, metrics, and logs. Glossary entry: https://moda.dev/glossary#opentelemetry ## Reasoning loops A behavioral failure where the agent retries the same failed action, oscillates between answers, or fails to converge. Glossary entry: https://moda.dev/glossary#reasoning-loops ## Segment A topic-coherent slice of a conversation, identified by cosine drift between consecutive turn-pair embeddings. Glossary entry: https://moda.dev/glossary#segment ## Self-improving agent An AI agent that updates its harness — prompts, tools, workflows, context, memory, evals — from production signal, in a latent space outside the model weights so improvements apply across whichever model the harness mounts and can adapt to each user. Glossary entry: https://moda.dev/glossary#self-improving-agent ## Tool misuse A behavioral failure where the agent calls the right tool with wrong arguments — or the wrong tool — while still returning a plausible response. Glossary entry: https://moda.dev/glossary#tool-misuse ## Tool call failure An explicit failure of a tool call — timeout, error response, or invalid arguments — surfaced with error subtypes and trends. Glossary entry: https://moda.dev/glossary#tool-call-failure ## Zero-config ingest Three lines of SDK code plus OpenTelemetry-native intake — works with any LLM provider, no lock-in. Glossary entry: https://moda.dev/glossary#zero-config-ingest ## Moda skills for AI agents Moda is a harness engineering platform: it turns production traces into verified improvements for the agent harness — prompts, tools, skills, evals, memory. Machine-readable artifacts for agents working with Moda: - Claude Code integration skill: https://moda.dev/skills/claude-code.md - Moda CLI skill: https://moda.dev/skills/moda-cli.md - Node.js SDK integration skill: https://moda.dev/skills/sdk-node.md - Python SDK integration skill: https://moda.dev/skills/sdk-python.md - Agent skills index: https://moda.dev/.well-known/agent-skills/index.json - LLM reference: https://moda.dev/llms.txt ## See it on your traffic Book a demo: https://cal.com/team/moda/demo-meeting?overlayCalendar=true