Skip to main content
p0’s AI uses tools to interact with your codebase. Every tool call is visible with real-time output, giving you complete transparency into what the AI is doing.

What are tools

Tools are capabilities the AI uses to accomplish tasks — reading files, running commands, editing code, searching the web. Each tool call shows:
  • Inputs: The parameters passed to the tool
  • Outputs: The results returned by the tool
  • Status: Whether the tool succeeded or failed
Every action the AI takes goes through a tool. Nothing happens invisibly.

Real-time visualization

p0 visualizes tool execution with animated status indicators:
  • Pending: Shimmer animation while the tool is queued
  • Running: Animated purple border during execution
  • Success: Green checkmark when the tool completes successfully
  • Error: Red indicator with error details if the tool fails
This lifecycle visualization helps you understand exactly what’s happening at any moment.
Tool execution panel showing real-time status

Tool categories

p0 provides 27+ tool types organized into categories:

File operations

Read, Edit, Write, Grep, Glob for working with your codebase

Execution

Bash commands with streaming output and exit codes

Web

WebSearch and WebFetch for external documentation and APIs

Planning

Task (subagents), TodoWrite, PlanWrite, and Thinking for complex workflows

Permission model

p0 uses a permission model to control tool access:
  • Plan mode: Read-only tools only (Read, Grep, Glob, WebFetch). The AI can explore but cannot make changes.
  • Agent mode: All tools enabled. The AI can edit files, run commands, and make changes to your codebase.
This ensures you maintain control over when the AI can modify your project.
You can switch between Plan and Agent modes at any time using /plan or /agent in the chat.

Explore each tool