Glossary

Our glossary defines the key terms and concepts behind agentic coding, testing, and the modern software development lifecycle.

#1
ADLC (Agentic Development Lifecycle)
ADLC, or Agentic Development Lifecycle, is an AI-enabled SDLC where agents take part in planning, coding, review, testing, documentation, and remediation. Learn how it works and how it differs from the traditional SDLC.
2 min read
#2
Agent Handoff
An agent handoff is the transfer of context, task state, and outputs from one agent to the next. Learn what crosses the boundary, the failure modes, and how it differs from an agent-as-tool call.
3 min read
#3
Agentic Orchestration
Agentic orchestration is how multiple agents and workflows are coordinated to deliver work across the lifecycle. Learn the patterns, the modular event-driven model, and how it differs from a single workflow.
2 min read
#4
Agentic SDLC
The Agentic SDLC runs software delivery as automated, policy-driven workflows where autonomous agents do lifecycle work and humans stay in control. Learn how it works and how it differs from AI-assisted development.
2 min read
#5
Agentic Workflow
An agentic workflow is a defined, repeatable sequence of agent and human steps, started by an event, that automates one software task end to end. Learn how it works and how it differs from manual agent use.
2 min read
#6
Audit Trail
An audit trail is the chronological, tamper-evident record of every action an AI agent takes, plus the context and reasoning behind it. Learn what it captures and how it differs from application logging and monitoring.
3 min read
#7
Automated Code Review
Automated code review is an agentic workflow that reviews every pull request against your standards and posts actionable findings before a human approves the merge. Learn how it differs from linters.
3 min read
#8
Autonomous Agent
An autonomous agent is a goal-directed AI worker that plans and carries out multi-step tasks by reasoning and calling tools. Learn how its perceive-reason-act loop works and how it differs from a copilot.
3 min read
#9
Control Plane
A control plane is the coordination layer that decides what AI agents do, what they can access, and when their work advances. Learn how it differs from the data plane and how it governs agents across the SDLC.
2 min read
#10
Ephemeral Development Environment
An ephemeral development environment is a temporary workspace created on demand for a single task and torn down when the work is done. Learn how it works and how it differs from a persistent environment and an isolated execution environment.
3 min read
#11
Fix CI
Fix CI is an agentic workflow that detects a failing CI run, diagnoses the cause from the logs, and fixes the code on the branch for a human to merge. Learn how it works and where its limits are.
3 min read
#12
Governance Gates
Governance gates are approval or validation checkpoints that control when an AI agent can move work forward in the SDLC. Learn how automated, human, and hybrid gates work, and how they differ from CI/CD quality gates and guardrails.
2 min read
#13
Guardrails
Guardrails are the predefined boundaries that constrain what an AI agent can do, access, or change. Learn the two families of guardrail and how they differ from policy enforcement and governance gates.
3 min read
#14
Human-in-the-Loop
Human-in-the-loop (HITL) is the model where a person participates at defined decision points so agents advance only with human judgment. Learn the in, on, and out-of-the-loop spectrum and how to tune it to risk.
3 min read
#15
Isolated Execution Environment
An isolated execution environment is a secure workspace where AI agents can inspect, run, or modify code without exposing production systems or data. Learn how it works and how it differs from a sandbox and a production environment.
2 min read
#16
Levels of Autonomy
Levels of autonomy are a maturity scale for how much an AI agent decides and executes versus how much a human directs and reviews. Learn the ladder from assist to fully autonomous, why teams plateau at level 3, and what it takes to climb.
2 min read
#17
Multi-Agent System
A multi-agent system is a model where several specialized AI agents work together on a problem too large for one agent. Learn the architecture patterns, why specialization beats one generalist, and how it differs from orchestration.
2 min read
#18
Playbook
A playbook is a reusable template that packages a proven agentic workflow, its steps, prompts, and triggers, so a team can import, customize, and run it. Learn how it differs from a workflow and a runbook.
2 min read
#19
Policy Enforcement
Policy enforcement is the mechanism that checks each AI agent action against your rules and allows or blocks it at execution time. Learn how policy-as-code works and how it differs from guardrails and governance gates.
2 min read
#20
Root Cause Analysis
Root cause analysis is the process of tracing a failure to its underlying cause instead of treating the symptom. Learn the methods, the complex-systems caveat, and how agents automate RCA across the SDLC.
3 min read
#21
Task Decomposition
Task decomposition is how an agent breaks a complex goal into smaller, ordered subtasks it can actually execute. Learn the approaches, dependencies, and how it differs from chain-of-thought reasoning.
2 min read
#22
Trigger
A trigger is the rule that automatically starts an agentic workflow when a specific engineering event occurs and its conditions are met. Learn the event, schedule, condition, and manual types, and how a trigger differs from an event and a webhook.
2 min read

“With Overcut, we finally standardized our workflows and embedded AI at every critical checkpoint, elevating code quality and minimizing operational risk.”

Asaf David, VP R&D, Guidde

+20%

hours saved per week with overcut

80%

faster PR reviews

Glossary

ADLC (Agentic Development Lifecycle)

Agent Handoff

Agentic Orchestration

Agentic SDLC

Agentic Workflow

Audit Trail

Automated Code Review

Autonomous Agent

Control Plane

Ephemeral Development Environment

Fix CI

Governance Gates

Guardrails

Human-in-the-Loop

Isolated Execution Environment

Levels of Autonomy

Multi-Agent System

Playbook

Policy Enforcement

Root Cause Analysis

Task Decomposition

Trigger