Contrarian Researcher — Development agent for Claude Code
Finds counter-evidence, challenges consensus, surfaces dissenting positions and null results that the mainstream framing buries.
How to install Contrarian Researcher
Installs to ~/.claude/agents/devergehq-locus-contrarian-researcher.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/devergehq/locus/HEAD/agents/contrarian-researcher.md -o ~/.claude/agents/devergehq-locus-contrarian-researcher.md Restart Claude Code, or start a new session, for it to be picked up.
What Contrarian Researcher does
id: contrarian-researcher name: Contrarian Researcher description: Finds counter-evidence, challenges consensus, surfaces dissenting positions and null results that the mainstream framing buries. model_preference: opus
Contrarian Researcher
**Role:** Finds counter-evidence, challenges consensus, surfaces dissenting positions and null results that the mainstream framing buries.
Stance (composed from traits)
- research — full research discipline
- contrarian — actively see
Alternatives in Development
- Key Exports From Each Phase — for summary in .planning/phases//-SUMMARY.md; do echo "=== $summary ===" grep -A 10 "Key Files\ Exports\ Provi 38.5k ★
- Tend — Tend the Allium garden 3.7k ★
- CORS Hunter — CORS Misconfiguration specialist (H1 #58) 812 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Independent Thinker
Contrarian analyst who challenges assumptions with evidence, presents alternative viewpoints, and declares unc
Reverse Salient Agent
Surfaces reverse-salient findings (Engine 1 Act 1) as F.0 Decision Gates with persona-aware framing. Sibling t
Selfish Original Thinking
Develop the user’s own ideas, frameworks, distinctions, hypotheses, and contrarian perspectives from lived exp
Reviewer2
Adversarial scientific reviewer (Reviewer 2). Spawned as subagent to DESTROY weak claims. Assumes every claim
Debate Tribunal
Multi-agent structured debate protocol where 3-5 persona-driven debaters argue positions through claim-evidenc
Op Debater
Read-only debate seat — holds ONE position across rounds, engages rival positions on their strongest form, and
Related Skills
Roast
Enforces Law 1 (Research Before Executing) of the 7 Laws of AI Agent Discipline. Convene a 5-persona adversari
Plan Ceo Review
CEO challenges the mission's scope and framing before any code
Claw Consensus
Run AI agents across machines without state drift: consensus protocol (propose/counter/accept/commit), dual-ra