New to 100xprompt? Start with Agent, Session, Tool, and Permission mode - they anchor almost everything else.
A - C
| Term | Definition |
|---|---|
| Agent | A focused operator that handles a task, with its own instructions, model, and set of allowed tools. You always talk to one primary agent per session. See Agents & Subagents. |
| Background task | Work that runs asynchronously - a long build, a watcher, or a delegated job - so your session stays responsive while it continues. See Background Tasks. |
| Command | A reusable prompt you trigger with a slash, like /init. Commands package a repeatable instruction so you can run it in one keystroke. See Commands. |
| Context | The files, history, rules, and memory the agent draws on to answer your request. Managing context well is what keeps long sessions accurate. See Memory & Context. |
M - P
| Term | Definition |
|---|---|
| Marketplace | A catalog you publish and install plugins from, so a whole team can share the same skills, commands, agents, and MCP servers. See Marketplace. |
| MCP | Model Context Protocol - the standard for connecting 100xprompt to external systems like issue trackers, docs, databases, and browsers. A connected MCP server’s capabilities appear as regular tools. See Connect Tools & Data with MCP. |
| Memory | Long-term knowledge 100xprompt carries across sessions - preferences, decisions, and hard-won project context - so you don’t repeat yourself. See Memory & Context. |
| Model | The 100xprompt model driving a session - Flash for fast, lightweight work or Pro for deep reasoning and large refactors. Switch it mid-session or set a default. See Models. |
| Permission mode | The stance a session takes toward consequential actions: asking before changes (default), auto-accepting edits (acceptEdits), read-only investigation (plan), or running freely (auto). See Permissions. |
| Plugin | An installable bundle of skills, commands, agents, and MCP servers packaged together, so capabilities install and update as a unit. See Plugins. |
| Provider | The service that runs a model. 100xprompt serves its own models - Flash and Pro - through its hosted service. See Models. |
R - T
| Term | Definition |
|---|---|
| Rules / AGENTS.md | Project instructions you write in an AGENTS.md file that 100xprompt reads and follows - conventions, do’s and don’ts, and how your codebase works. See Rules. |
| Schedule | A recurring or timed run that fires automatically, so routine work happens without you starting it by hand. See Scheduling. |
| Session | A single continuous conversation with 100xprompt, including its history, context, and results. Sessions can be continued, shared, or resumed later. See How It Works. |
| Share link | A URL that opens a shared session in the browser for others to view. Create one with the share action or --share, and control default behavior with the share setting. See Sharing. |
| Skill | A packaged, repeatable procedure - instructions plus any helper files - that 100xprompt can invoke when it’s relevant to your request. See Skills. |
| Subagent | An agent spun up by another agent to handle a self-contained chunk of work in parallel - in its own isolated context, reporting back a concise result. See Agents & Subagents. |
| Tool | A concrete action an agent can take - reading a file, editing code, running a command, or a capability provided by an MCP server. Tools are what turn a plan into changes. See Tools. |
How the terms fit together
Every term above belongs to one of five families. This map shows how the core concepts branch out from the product, so you can see the shape of the whole vocabulary at a glance.Related
FAQ
Quick answers to the most common questions about 100xprompt.
Capabilities Overview
See what these concepts unlock, end to end.
Extend 100xprompt
Skills, commands, plugins, MCP, and marketplaces in depth.
Troubleshooting
Diagnose and fix common issues fast.