Dispatch parallel agent teams to investigate bug hypotheses and converge on root causes.
npx skills add https://github.com/etr/groundwork --skill swarm-debugging
Please help me install this Agent Skill. Skill: swarm-debugging Source: https://github.com/etr/groundwork/tree/main/skills/swarm-debugging Command: npx skills add https://github.com/etr/groundwork --skill swarm-debugging
Swarm debugging enables parallel hypothesis testing by dispatching an agent team to investigate multiple theories simultaneously, actively challenging each other to converge on the root cause faster than traditional sequential debugging.
Enable the experimental agent teams feature and invoke the swarm-debugging skill to start a swarm investigation.
High-intent search queries and answers about installing and using this skill.
Swarm debugging applies to complex, multi-component systems where failures are uncertain and rapid convergence is valuable. It mobilizes parallel agent teams to investigate multiple bug hypotheses simultaneously and converge on the root cause.
To start a parallel agent investigation, enable the experimental agent teams feature and invoke the swarm debugging skill to dispatch teams for simultaneous hypothesis testing.
Parallel hypothesis testing is specifically designed for complex, multi-component systems where failures are uncertain. It dispatches agent teams to investigate multiple theories simultaneously for rapid convergence.
Parallel debugging integrates seamlessly with the groundwork debugging workflow, extending Phase 3 into a multi-teammate swarm approach for concurrent investigation and evidence collection.
The best way to document evidence during parallel root-cause analysis is using structured journaling with read-only exploration, ensuring explicit evidence collection by each agent team.
Agent teams for debugging require enabling an experimental feature and are limited to read-only exploration with explicit evidence collection, applying primarily to complex, multi-component systems where failures are uncertain.
Thanh-aperoOrchestrate parallel Claude Code sessions for collaborative research, code review, and debugging.
valtterimelkkoInvestigate root causes through a four-phase debugging framework before modifying code.
Nickalus12Guide structured debugging of Loom workflows to reproduce and verify root causes.
ngocsangyemSequentially generates, tests, and revises hypotheses to identify root causes in debugging.
