Koneetiv Academy

The glossary of AI, Claude and Claude Code

89 clear definitions, from the terminal to autonomous agents. The lexicon of the Koneetiv Academy course, freely available.

89 definitions6 categories

Terminal

23 definitions
pwd
Shows the folder you are currently in (print working directory).pwd
ls
Lists the files and folders in the current folder.ls
ls -la
Detailed listing, hidden files included (those whose name begins with a dot).ls -la
cd
Enters a folder (change directory).cd "Mon Dossier"
cd ..
Goes back up to the parent folder.cd ..
mkdir
Creates a folder.mkdir "nouveau-projet"
cp
Copies a file or a folder.cp rapport.docx rapport-v2.docx
mv
Moves or renames a file.mv ancien.txt nouveau.txt
rm
Permanently deletes, with no trash bin. An ideal candidate for a deny rule.rm fichier.txt
cat
Displays the contents of a file in the terminal.cat notes.txt
open
(macOS) Opens a file or a folder with the default application.open index.html
grep
Searches for text within files.grep "budget" *.txt
clear
Clears the terminal display (the command history remains).
history
Shows the most recent commands you have typed.history
| (pipe)
Sends the output of one command to the next command.history | grep claude
> and >>
> writes the output to a file (overwriting it); >> appends it to the end.ls > liste.txt
~
Shortcut to your home folder.cd ~/Desktop
. and ..
The current folder (.) and the parent folder (..).
Absolute / relative path
An absolute path starts from the root of the disk (/Users/…); a relative path starts from the folder you are in (folder/file.txt).
Quotation marks
Required around a path containing spaces, otherwise the terminal sees several separate arguments.cd "France API 26"
sudo
Runs a command with administrator rights; the terminal asks for your password.
curl
Downloads the content of a web address. It is what installs Claude Code.curl -fsSL https://claude.ai/install.sh | bash
claude
Launches Claude Code in the current folder.claude

Claude Code Commands

15 definitions
/init
Analyzes the project and generates an initial CLAUDE.md.
/clear
Clears the conversation to start over with a clean context. Rule of thumb: one conversation = one task.
/compact
Summarizes the ongoing conversation to free up context without losing everything.
/model
Switches the model mid-session (Opus, Sonnet, Haiku).
/rewind
Opens the rewind menu: restore the code, the conversation, or both. Equivalent: pressing Esc twice.
/help
Lists the available commands.
/permissions
Manages allow rules (no confirmation) and deny rules (always blocked).
/agents
Manages custom agents (roles, allowed tools, instructions).
/mcp
Manages MCP servers: connector status, authentication.
/memory
Opens the memory files, including CLAUDE.md.
/config
Opens the Claude Code settings.
/review
Requests a review of the current code.
/cost
Displays the session's usage.
Esc
Interrupts Claude mid-action without breaking the session. You can then redirect it.
Esc Esc
Double press: opens the rewind menu (equivalent to /rewind).

Hooks: all events

14 definitions
Hook
An automatic command run by the system at a specific point in Claude's work. Unlike an instruction, a hook cannot be forgotten: it fires every time.
PreToolUse
Fires just before a tool runs. It can block the action: this is the guardrail event (for example, forbidding any deletion).
PostToolUse
Fires just after a tool runs. Ideal for follow-up automations (for example, formatting the code after each file write).
UserPromptSubmit
Fires when you send a message, before it is processed. It lets you add context or vet the request.
Stop
Fires when Claude finishes its response. Common use: a notification or an end-of-task sound.
SubagentStop
Fires when a subagent finishes its work.
SessionStart
Fires when a session opens: loading context, preparing the environment.
SessionEnd
Fires when a session closes: cleanup, logging.
Notification
Fires when Claude Code emits a notification (for example, a pending permission request).
PreCompact
Fires just before the context is compacted: the last moment to save what matters.
matcher
A filter that limits a hook to a specific tool (for example, Write to react only to file writes)."matcher": "Write"
command (hook)
The shell command run when the hook's event fires.
Exit code 2
If a PreToolUse hook's command exits with code 2, Claude's action is blocked.
timeout (hook)
The maximum time allowed to a hook before it is interrupted.

Claude & AI Concepts

20 definitions
Token
A piece of a word, the unit of measurement and billing. You count what Claude reads (input) and what it writes (output).
Context
The working memory of a conversation: everything that has been said, read, and produced.
Context window
The maximum size of the context. When it fills up, Claude automatically summarizes the older content to keep going.
Compaction
The summary of the context (automatic or via /compact) that lets you continue a long session.
Prompt
The instruction you give Claude. Four components: context, a precise task, constraints, and the expected format.
Model
The engine that answers. Opus: the most powerful. Sonnet: the balance. Haiku: fast and economical.
Agentic
Able to act using tools (reading, writing, running commands) and not just to answer.
Subagent
An instance of Claude launched by Claude for a subtask, in an isolated context. It returns only its conclusion.
Orchestration
Coordination of several agents: in parallel (fan-out), in a chain (pipeline), or in cross-checking (adversarial verification).
Fan-out
Splitting a large task into pieces handled in parallel by several agents.
Pipeline
Chaining steps together: each item moves forward independently from step to step.
Adversarial verification
One agent produces, another tries to refute: the conclusions that hold up are reliable.
Checkpoint
A restore point created automatically before changes. It is the basis of /rewind.
Diff
The view of the proposed changes: green = added, red = removed. Always review it before accepting.
Skill
A folder that packages a piece of know-how (a SKILL.md file plus resources). Claude activates it on its own when the situation matches its description.
Plugin
A ready-to-use pack for a given role: connectors plus skills.
Artifact
Self-contained content created by Claude on claude.ai (a page, tool, or document), displayed in a dedicated panel and shareable by link.
Persistent storage
An artifact's ability to save data durably, with a separate space for each user.
Magic link
Passwordless sign-in by email: you click the link you receive and the session opens.
Prompt injection
Hidden instructions placed inside external content (an email, a web page, a document) to manipulate the AI. The countermeasure: review any action triggered by external content before approving it.

MCP & Connectors

7 definitions
MCP
Model Context Protocol: the standard that plugs Claude into external services.
MCP Server / Connector
Two names for the same thing: a link to a service (Gmail, Calendar, Figma, Slack, Notion, GitHub, etc.).
Tool (MCP)
An action exposed by a connector.create_event, send_message
Resource (MCP)
A piece of data a connector makes available for Claude to read.
Prompt (MCP)
A reusable instruction template provided by a connector.
Authentication (OAuth)
Access authorization granted once through a secure window. Revocable at any time from the service concerned.
Least privilege
Security rule: connect only the connectors you need, and only to the accounts you need.

Files & configuration

10 definitions
CLAUDE.md
The project's permanent instruction sheet, reread automatically at the start of every session: description, conventions, useful commands, prohibitions.
settings.json
The central configuration file: permissions, hooks, environment variables, preferences.
Settings scopes
Global (the whole machine), project (shared with the team), local (personal to the project). The most specific one wins.
allow
The list of actions Claude runs without asking you for confirmation.
deny
The list of actions that are absolutely forbidden. The safeguard: perfect for protecting a sensitive folder.
Environment variable
A reusable setting (an API key, an option) stored outside the code, declared notably via settings.json.
.claude/
A project's customization folder: settings, commands, agents, skills.
git
The professional code-backup system: each commit is a complete snapshot of the project.
commit
A named backup of the project's state. A good reflex before a big change: 'make a commit'.
Branch
A parallel line of work in git: you experiment without touching the main version, then merge it back.

From definition to practice

This glossary is the lexicon of Koneetiv Academy. Take the free course to put these ideas into practice, a trial lesson with no account.

Explore the course