85 agentic tools audited for security incidents · Talk to the lab →
MotionticBook a call
Glossary

Agentic AI, defined

The Motiontic glossary defines the terms used in building AI agents and AI products, one term per page, each opening with a single-sentence definition.

All glossary

Glossary

A2A protocol

The open protocol for agents talking to other agents, now under the Linux Foundation.

Glossary

Agent Card (A2A)

The business card an A2A agent publishes so other agents can find and call it.

Glossary

Agent-computer interface (ACI)

The tools, docs and outputs an agent sees, designed as carefully as a UI.

Glossary

Agent framework

A library of agent building blocks: loop, tools, memory, orchestration and traces.

Glossary

Agent gateway

One control point for every call an agent makes.

Glossary

Agent graph

Nodes for model calls and code, edges for what runs next, and state that flows through.

Glossary

Agent handoff

One agent passes the conversation to another, which carries on from there.

Glossary

Agent harness

The loop, tools, context, permissions and state that sit around the model.

Glossary

Agentic AI

AI that acts toward a goal, from model-assisted workflows to fully model-directed agents.

Glossary

Agentic coding

Engineers delegate whole tasks to a coding agent and spend their time on direction and review.

Glossary

Agentic eagerness

The dial between an agent that over-explores and one that hands back too soon.

Glossary

Agentic search

Retrieval as a tool the agent calls, step by step, rather than a pipeline stage.

Glossary

Agentic workflow

Model calls on a path set in code: predictable, cheaper and easier to test than a free-running agent.

Glossary

Agent identity

Each agent gets its own credential, permissions and audit trail.

Glossary

Agent loop

Model call, tool call, result, repeat: the cycle at the centre of every agent.

Glossary

Agent memory

Session state for the task at hand, plus stored knowledge the agent recalls later.

Glossary

Agent observability

Step-by-step traces of what an agent did, why, and what it cost.

Glossary

AgentOps

DevOps and MLOps, adapted for systems that decide their own next step.

Glossary

Agent orchestration

The logic that decides what runs next, with what context, and who combines the results.

Glossary

Agent quality flywheel

Production failures feed the eval suite, and the suite gates the next release.

Glossary

Agent runtime

Where agents actually run: durable sessions, checkpoints, sandboxes and resume.

Glossary

Agent sandbox

Hard walls around where an agent can read, write and connect.

Tool explainer

Agent Skills

Folders of instructions, scripts and resources that an agent loads only when the task needs them.

Glossary

AGENTS.md

A README for coding agents: the file an agent reads first in a repo.

Glossary

Agent state

The facts about where a task stands, kept as data instead of buried in chat history.

Glossary

Agent teams

Several full agent sessions on one project, claiming tasks and talking to each other.

Glossary

AG-UI

The protocol between an agent backend and the app a person is using.

Glossary

AI agent

A model that picks its own next step and tool, in a loop, until the job is done.

Glossary

AI guardrails

Checks on what goes into a model, what comes out, and what an agent is allowed to do.

Glossary

AI hallucination

Confident output that is not true, and how production systems contain it.

Glossary

AI red teaming

Attacking an AI system on purpose, so the attacks that matter are found first.

Tool explainer

Amazon S3 Vectors

Vector storage and similarity search built into S3, priced like storage rather than a database.

Glossary

Approval fatigue

Too many approval prompts, and people stop reading them.

Glossary

Artifacts (AI agents)

Large data kept out of the prompt, or outputs that show a person what the agent did.

Glossary

Attention budget

Every token in context costs some of the model's focus, not just some of its window.

Glossary

Augmented LLM

A model plus retrieval, tools and memory: the unit every workflow and agent is built from.

Glossary

Background agent

An agent that works while nobody watches, and hands back a pull request.

Glossary

Blast radius (AI agents)

The worst an agent could do, measured by what it can reach.

Glossary

BM25

The keyword ranking function behind most full-text search, and half of hybrid search.

Glossary

Capability eval

Hard tasks the agent fails today, kept to measure progress.

Glossary

Chain-of-thought prompting

Asking the model to show its working, now mostly built into the model itself.

Glossary

Chunking (RAG)

Splitting documents into passages before indexing, which quietly sets retrieval quality.

Glossary

Classifier-gated permissions (auto mode)

A model reviewer in place of a permission prompt, with published error rates.

Tool explainer

Claude Agent SDK

Claude Code's harness as a library: the same loop, tools and context management, in a team's own process.

Glossary

Coding agent

An agent with a shell and a repository: it reads, edits, runs the tests and tries again.

Glossary

ColPali

Searching PDFs as images, with one vector per patch instead of extracted text.

Glossary

Compaction

Summarise the old turns, keep working: the standard move for long agent runs.

Glossary

Computer use agent

Agents that work through the screen for systems with no API.

Glossary

Context anxiety

The model thinks it is running out of room, so it rushes to finish.

Glossary

Context editing

Clear the old tool results, keep the conversation, keep the agent running.

Glossary

Context engineering

Curating everything in the model's view, not just the wording of the prompt.

Glossary

Context reset

Throw away the window, keep the handoff, start the next session clean.

Glossary

Context rot

Why agents get worse the longer they run, and what keeps context clean.

Glossary

Contextual retrieval

Give every chunk a line of context before it is indexed, so it can still be found.

Glossary

Context window

The model's working memory for one request, measured in tokens, and finite.

Glossary

Cosine similarity

The angle between two embeddings, and the default way to compare them.

Tool explainer

Deep Agents

LangChain's batteries-included, model-neutral harness for long-running agents.

Glossary

Delegated approval

An approval gate that can wait days without costing anything.

Glossary

Deployment overhang

Models could do more on their own than they are currently allowed to.

Glossary

DiskANN

Graph vector search that lives on SSD, not in RAM.

Glossary

"Done when" criteria

Completion written down as checks before the agent starts, not judged after.

Glossary

Dual LLM pattern (privileged and quarantined LLMs, CaMeL)

Keep the model that acts away from the text an attacker controls.

Glossary

Durable execution

Runs that survive crashes and pauses because every finished step is recorded.

Glossary

Durable project memory

The spec, plan and status files a long agent run is anchored to.

Glossary

Embeddings

Meaning turned into vectors, so similar things can be found by distance.

Glossary

Eval

Repeatable, scored tests that tell a team whether an AI change made things better or worse.

Glossary

Eval awareness

When the model notices the test, the score stops meaning what it should.

Glossary

Eval harness

The machinery that runs evals, separate from the agent it tests.

Glossary

Eval saturation

When every score is near 100% and the eval stops telling anyone anything.

Glossary

Execution derailment

When an agent loop loses the thread as its context fills up.

Glossary

Extended thinking

Claude reasoning in thinking blocks before it answers, now steered by effort.

Glossary

Few-shot prompting

Teaching a model the task by showing a few examples, not by retraining it.

Glossary

Filtered vector search

Vector search with a WHERE clause, which is harder than it looks.

Glossary

Fine-tuning

Changing the model itself with examples, rather than changing what it is shown.

Glossary

Generator-critic loop

One model drafts, another grades, and the loop runs until the grade passes.

Glossary

Golden dataset

The reviewed, versioned test cases an eval trusts as ground truth.

Glossary

Golden principles

Repo rules that background agents enforce continuously, like garbage collection for code.

Tool explainer

Google ADK

Google's open-source agent framework, now with graph workflows, in five languages.

Glossary

Grader (AI evals)

The part of an eval that decides pass or fail.

Glossary

GraphRAG

RAG over a model-built knowledge graph, for questions about a whole corpus.

Glossary

Harness engineering

Designing the environment an agent works in, so its output can be checked and shipped.

Glossary

HNSW

The layered graph index behind most vector search.

Glossary

Human in the loop

Approval gates where a person decides, placed by risk rather than on every step.

Glossary

Hybrid search

Keyword ranking and vector similarity, fused into one result list.

Glossary

Infrastructure noise (evals)

Score changes that come from the machine, not the model.

Glossary

Initializer agent

The one-off first session that sets up everything later agent sessions rely on.

Glossary

IVF index

Vector search by clusters: pick the nearest few, scan only those.

Glossary

Just-in-time context

Keep the address in context, fetch the contents when the step needs them.

Glossary

Large language model (LLM)

A next-token predictor trained at scale, and the engine inside every AI agent.

Glossary

Late interaction (ColBERT)

Many vectors per passage, matched token by token: finer retrieval, more storage.

Glossary

LLM as a judge

A model grading model output against a rubric, checked against human labels.

Glossary

Long-running agent

Agents that work for hours or days, and the state that lets them pick up where they stopped.

Glossary

Long-term memory (AI agents)

What an agent keeps after the conversation ends, and the rules for writing and recalling it.

Glossary

Manager pattern (agents as tools)

One agent stays in charge and calls specialist agents the way it would call any other tool.

Glossary

Matryoshka embeddings

Embeddings that can be cut short, like nested dolls, without starting over.

Glossary

MCP server

The program that wraps a system so any MCP-compatible agent can use it.

Glossary

Memory tool

A file directory Claude can read and write across sessions, stored where the developer chooses.

Glossary

Meta-harness (brain vs hands)

Stable interfaces that split reasoning, execution and memory so each can fail alone.

Tool explainer

Microsoft Agent Framework

Microsoft's one forward path for agents, replacing Semantic Kernel and AutoGen.

Glossary

Model Context Protocol (MCP)

The open standard for plugging tools and data into AI agents.

Glossary

Model-native harness

A harness shipped by the model's own maker, tuned to how that model works.

Glossary

Multi-agent system

Several agents splitting one job, with some way of coordinating who does what.

Glossary

Multimodal RAG

RAG over scans, slides, charts, audio and video, not only extracted text.

Glossary

Narrative casting

Show the next agent what another agent did, labelled as someone else's work.

Glossary

No-code agent builder

Build an agent by configuring it in a visual tool instead of writing code.

Glossary

Object-storage vector database

Vector search built on S3-style storage, with a cache for the hot data.

Glossary

Online evals

Grading live traffic in production, not only test sets before release.

Tool explainer

OpenAI Agents SDK

OpenAI's code-first agent framework: a small set of primitives that run in a team's own app.

Glossary

OpenTelemetry GenAI semantic conventions

Standard gen_ai.* names for agent and model telemetry, still at Development status.

Glossary

Orchestrator-worker pattern

A lead agent plans, workers run in parallel with their own context, the lead merges.

Glossary

Parallelization (agent pattern)

Several model calls at once, on separate parts of a task or as repeated votes, merged in code.

Glossary

Parallel tool calls

Several tool calls in one model turn, run together to save round trips.

Glossary

pass@k and pass^k

Can the agent do it at all, versus does it do it every time.

Glossary

Plan-and-execute agent

Plan the whole task first, then execute the steps, and replan only when something breaks.

Glossary

Programmatic tool calling

The model writes a script that calls the tools, so raw results never fill its context.

Glossary

Progressive disclosure

Show the menu first, load the dish only when it is ordered.

Glossary

Prompt caching

Paying less for the part of the prompt that does not change between calls.

Glossary

Prompt chaining

A fixed sequence of model calls, each one easier than the whole task, with checks in between.

Glossary

Prompt injection

Untrusted text that hijacks a model's instructions, and how agents are built to survive it.

Glossary

Query rewriting

Fixing the question before searching, so retrieval has something to match.

Glossary

ReAct agent

Reason, act, observe, repeat: the pattern most agent loops descend from.

Glossary

Reasoning model

A model that thinks before it answers, at a price in tokens and time.

Glossary

Recall@k

Did the right passages make the shortlist? The first number to check in RAG.

Glossary

Reciprocal rank fusion (RRF)

Merging keyword and vector results by rank, not by score.

Glossary

Regression eval

The suite that catches what a change broke before users do.

Glossary

Reranker

A second pass that rereads the shortlist and puts the right passages first.

Glossary

Retrieval-augmented generation (RAG)

Fetch the right passages at query time, then answer from them with citations.

Glossary

Right altitude (system prompts)

Specific enough to steer, general enough not to break: the target for a system prompt.

Glossary

Routing (agent pattern)

Classify first, then send each input to the branch built for it.

Glossary

Self-reflection (agents)

An agent critiques its own attempt in words and tries again with that critique in hand.

Glossary

Semantic cache

Reuse an answer when a new question means the same as an old one.

Glossary

Session (AI agents)

The durable log of a run, kept outside the window, that every prompt is built from.

Glossary

Source-sink analysis

Find where attacks get in and what they could reach, then cut the path.

Glossary

Sparse vectors (SPLADE)

Learned keyword weights: the middle ground between BM25 and dense embeddings.

Glossary

Spec-driven development

Write down what the software must do first, then let the coding agent plan and build against it.

Glossary

Sprint contract

Generator and evaluator agree on done before any code is written.

Glossary

Structured note-taking

The agent keeps its own notebook outside the context window, in a format it can trust.

Glossary

Structured outputs

JSON that matches the schema because the decoder allows nothing else.

Glossary

Subagent

A helper agent with a clean context that hands back a short answer.

Glossary

Subagent sprawl

Too many overlapping subagents, each owned by a different team and none by the whole.

Glossary

System prompt

The standing instructions that frame every turn of a conversation or agent run.

Glossary

Task, trial and transcript (eval vocabulary)

The basic units every agent eval is built from.

Glossary

Temperature (LLM sampling)

The randomness knob on token sampling, and why the newest models dropped it.

Glossary

Think tool

A tool that does nothing, so the model has somewhere to stop and think between actions.

Glossary

Token (LLM)

The unit models read, write and bill in: a word or a piece of one.

Glossary

Tool calling

How a model asks software to do something: a named tool, typed arguments, a result.

Glossary

Tool preamble

The plain-language plan and progress notes a model writes around its tool calls.

Glossary

Tool risk rating

Sorting an agent's tools by how much damage a bad call could do.

Glossary

Tool search

Load tools when they are needed, not all at once before the work starts.

Glossary

Tool use examples

Sample calls on a tool definition that teach usage a schema cannot.

Glossary

Trace grading

Scoring a whole recorded agent run, step by step.

Glossary

Trajectory evaluation

Grading the route an agent took, not just where it ended up.

Glossary

Trimming vs summarization

Drop the old turns or compress them: the two basic short-term memory tactics.

Tool explainer

turbopuffer

Vector and full-text search built on object storage, priced for very large and many-tenant corpora.

Glossary

Vector database

Storage and nearest-neighbour search for embeddings, now in many shapes.

Glossary

Vector quantization

Fewer bits per dimension: smaller, cheaper indexes at a small cost in recall.

Glossary

Vibe coding

Prompt, accept, run, repeat: building by results without reading the code.

Glossary

Working context

The per-call view of the agent's state, compiled fresh for each model call.

Running agents that have to hold up?

Describe the agent, the systems it touches and what it must never get wrong. The first call is with a lab engineer, who will say plainly whether it is ready for production, and whether Motiontic is the right team.