Summarize the latest discussion fragment into a structured state snapshot.
npx skills add https://github.com/F010110/pbl-multi-agent --skill pbl-summarizer
Please help me install this Agent Skill. Skill: pbl-summarizer Source: https://github.com/F010110/pbl-multi-agent/tree/main/.opencode/skills/pbl-summarizer Command: npx skills add https://github.com/F010110/pbl-multi-agent --skill pbl-summarizer
The summarizer compresses only the most recent PBL discussion fragment to help the orchestrator recover the current discussion state without assuming full history.
Summarize only the most recent fragment of the discussion to help restore state and highlight key positions, disagreements, and next actions.
High-intent search queries and answers about installing and using this skill.
To recover the state of a multi-agent discussion, summarize only the latest PBL discussion fragment. This extracts each agent's current stance, disagreements, and open questions into a concise snapshot for next-step planning.
The best way to summarize multi-agent disagreements for an orchestrator is to condense the most recent PBL discussion fragment. This identifies main points of disagreement and unresolved questions to produce a structured state snapshot.
You extract individual agent stances from a PBL discussion fragment by processing the latest text through a summarizer. It isolates each agent's current position and open questions, outputting a structured summary for immediate orchestrator consumption.
Yes, you can use a discussion state summary to determine next steps in a multi-agent workflow. The summarizer outputs a concise state snapshot that highlights agent positions and unresolved questions, directly supporting next-step planning.
You need to summarize only the latest fragment of a multi-agent discussion when assuming full history is not possible. This state-recovery approach restores the current discussion context by focusing on recent agent positions and disagreements.
JmarquesGenerate and accumulate raw writing fragments through conversational user interviews.
ramonclaudioArchive Claude session state and health checks to the .handoff directory.
hainamchungRoute deployment tasks to the Agent Assistant workflow via /deploy in Codex-based agents.
lpjhelderDisplay agent session flow traces with chronological timelines and summaries.
