Orchestrate asynchronous sidecar sessions monitoring source session events.
npx skills add https://github.com/filipexyz/ravi --skill observers-filipexyz
Please help me install this Agent Skill. Skill: observers Source: https://github.com/filipexyz/ravi/tree/main/src/plugins/internal/ravi-system/skills/observers Command: npx skills add https://github.com/filipexyz/ravi --skill observers-filipexyz
This Skill solves the challenge of maintaining visibility into agent sessions without polluting the primary conversation context or runtime permissions.
Use the observers skill to list all active observer bindings and explain why a specific rule might not be triggering for your current session.
High-intent search queries and answers about installing and using this skill.
You can monitor asynchronous AI agent events without context pollution by using isolated sidecar sessions that receive canonical events from source sessions. This decoupled event processing approach maintains visibility into agent workflows while keeping the main agent prompt clean and focused.
Automated rule-based intervention for multi-agent workflows is the process of triggering specialized observer agents based on granular criteria like task state, profile, or tag. It automatically performs status updates or quality checks whenever a task transitions, ensuring continuous monitoring without manual oversight.
You configure observer triggers for specific agent sessions by defining granular rules based on global, agent, session, task, profile, or tag criteria. These rules are evaluated against source session metadata to automatically activate sidecar monitoring when specified conditions are met.
Yes, event-driven monitoring requires the Ravi runtime environment to evaluate selectors and reconcile observer bindings against source session metadata. This environment orchestrates the asynchronous sidecar sessions that handle decoupled event processing for complex multi-agent workflows.
Your observer rule might not be triggering due to a mismatch between the configured selector criteria and the current source session metadata. You can use the observers skill to list all active observer bindings and explain why a specific rule is not matching your session's state.
You should use decoupled event processing for task status tracking in complex multi-agent workflows requiring automated quality checks or status updates. It is ideal when you need to keep the main agent prompt clean while maintaining strict visibility into task state transitions.
filipexyzManage project-level alignment and context for AI agent workflows in the Ravi runtime.
huachengConvert .target.md into an executable .plan.md with tests and notes.
delorenjPublish and consume Bloodbank events with RabbitMQ routing and Pydantic payloads.
nashgaoExecute tasks in isolated sub-agent environments with result-only returns.
