~ / leaderboard / subagent / refuter
Subagent · #746
refuter
Adversarial reviewer that tries to falsify a hypothesis, analysis or draft section of the working paper. Use before a hypothesis is drafted and after each analysis result.
4rigs use it
0%of all rigs
2.1kavg rig context tax
Get it
$ npx degit gwpicard/ai-build-kit/.claude/agents .claude/agents From the most-starred rig that ships it: gwpicard/ai-build-kit. Same-named subagents in different rigs may differ — review before use.
Often used together with
reviewer · 2builder · 2research · 1researcher · 1implement · 1triage · 1merge · 1second-opinion · 1clarify · 1maintain · 1
Rigs using refuter (4)
gwpicard/ai-build-kit
A workflow for building and running internal tools with a coding agent.
Skill Collector 2.9k tok · · Q29
ncoevoet/claude-workflow-kit
Evidence-first Claude Code workflow as a plugin: verification standards with claim-class proofs, a 7-phase methodology with adversarial spec review, subagent model-tier enforcement hooks, and the CodeGraph MCP.
Pragmatist 3.2k tok · · Q0
luonghongthuan/agentic-shortlink
Hands-on project for the 'Shipping software with AI agents, A to Z' series: Claude Code, Codex, CI/CD, multi-provider
Fort Knox 1.2k tok · · Q54
alex-casapu/agentic-ai-sme-succession
Agentic AI, owner retirement and the back office of small firms: replication package and adversarial AI research log (EU vs US)
Minimalist 876 tok · · Q5