pi-monitors

Configure agent-monitor and hooks-based automation for pi from JSON files.

1|Updated May 10, 2026

Search Tags:#automation #reminders #agent #lifecycle #hook #monitor #pi

Ready to run

npx skills add https://github.com/cuevasclemente/mypi --skill pi-monitors

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: pi-monitors
Source: https://github.com/cuevasclemente/mypi/tree/main/skills/pi-monitors
Command: npx skills add https://github.com/cuevasclemente/mypi --skill pi-monitors

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Automates the setup of agent monitors and lifecycle hooks for pi, enabling proactive reminders and journaling prompts.

Core Features & Use Cases

Monitors evaluate agent output after each turn using a lightweight model to inject reminders when milestones are detected.
Hooks inject reminders at lifecycle points (after tools, every N turns, session start) without requiring LLM calls.
Both systems read their configuration from JSON files and merge global and project-local settings.

Quick Start

Install the agent-monitor and hooks extensions to ~/.pi/agent/extensions and place both global and project hooks.json files in the appropriate locations, then run /reload to apply the new monitors.

Frequently Asked Questions about pi-monitors

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I automate agent monitoring and lifecycle hooks for pi?▼

Automate agent monitoring and lifecycle hooks for pi by configuring JSON files to drive proactive reminders and journaling prompts after tool use or at session milestones. The system merges global and project-local settings to activate messaging rules.

What is the difference between agent monitors and hooks in pi?▼

Monitors evaluate agent output after each turn using a lightweight model to inject reminders when milestones are detected, while hooks inject reminders at lifecycle points like session start or after tools without requiring LLM calls.

How do I configure global and project-local hooks for pi?▼

Configure global and project-local hooks by placing JSON files in the appropriate locations, such as ~/.pi/agent/hooks.json and .pi/hooks.json. The system merges these configurations to validate monitor definitions and hook rules before activation.

Do I need extensions installed to use agent monitors and hooks with pi?▼

Yes, using agent monitors and hooks requires the agent-monitor and hooks extensions to be installed in your environment. You must install them to the extensions directory and run a reload command to apply the new monitors.

Can I inject reminders after tool use or at session milestones in pi?▼

Yes, you can inject reminders after tool use or at session milestones by configuring hook rules in JSON files. The system validates these rules and merges global and project-local settings to drive messaging without requiring LLM calls.

What are the limitations of using hooks-based automation for pi agents?▼

Hooks-based automation for pi agents requires the agent-monitor and hooks extensions to be installed and does not use LLM calls for injection. You must validate monitor definitions and hook rules in JSON files before activation to ensure proper merging.