Escalate risky coding tasks into multi-model debate and consensus gates with Claude Octopus
<p>Use Claude Octopus when ordinary Claude Code flows are not enough and a risky coding task needs adversarial multi-model review, disagreement surfacing, and a consensus gate before shipping.</p>
npx skills add agentskillexchange/skills --skill escalate-risky-coding-tasks-into-multi-model-debate-and-consensus-gates-with-claude-octopus
Claude Octopus is skill-shaped around one escalation workflow: move selected coding work from a normal Claude-native path into a stricter multi-model review lane. The agent fans work out to multiple models, compares their outputs, surfaces disagreements, and uses a consensus threshold before the operator treats the result as ready. The upstream project explicitly frames Octopus as the escalation path while ordinary tasks stay in standard Claude Code flows.
What this skill actually does
Invoke this instead of using Claude Code normally when the task is higher risk, the cost of a blind spot is meaningful, or you want adversarial review before shipping. The boundary prevents this from collapsing into a generic coding-agent listing: this entry is about the debate-and-consensus gate for risky tasks, not about every command, persona, or provider the project contains.