By snowmead
Advanced reasoning plugin integrating Chain of Verification (CoVe) and Atom of Thoughts (AoT) frameworks for rigorous, verifiable analysis
Decompose a complex problem into atomic sub-questions using Atom of Thoughts Graph Generator
Apply full UltraThink reasoning pipeline (AoT + CoVe) to analyze the current problem with maximum rigor
Verify a factual claim or answer a question with self-verification using the atomic solver
Use this agent to build a DAG structure of atomic questions from a complex problem. This agent creates ONLY the question graph - no solving. ## Examples: <example> Context: User asks a multi-part question requiring synthesis user: "How does React's reconciliation work and how does it compare to Vue?" assistant: "I'll use the aot-graph-generator agent to decompose this into atomic questions." </example> <example> Context: Planning a complex implementation user: "Help me build an authentication system with OAuth and rate limiting" assistant: "Let me build the question graph using the aot-graph-generator agent." </example>
Use this agent to contract existing atom questions with solved answers from dependencies. This agent reads answer.md from solved atoms and rewrites question.md for dependent atoms. ## Examples: <example> Context: A1 and A2 have been solved, need to contract A3's question assistant: "I'll use the aot-graph-maintainer agent to contract A3's question with the solved answers." </example>
Use this agent to evaluate answer quality across solved atoms at a level. This agent checks for coherence, contradictions, and completeness WITHOUT re-answering questions. ## When to use: * Rigor level is High-Stakes * After solving all atoms at a level * When additional quality assurance is needed beyond self-verification ## Examples: <example> Context: High-stakes rigor evaluation after Level 1 solving assistant: "I'll invoke the aot-judge to evaluate answer quality before proceeding." </example>
Use this agent to generate claims and verification questions for ONE atomic question. This agent does NOT fact-check or research - it only decomposes the question into verifiable claims. The orchestrator will spawn separate cove-verifier agents to research each verification question. ## Examples: <example> Context: Generating claims for an independent atomic question assistant: "I'll spawn cove-claim-qs to generate claims and verification questions." </example> <example> Context: Processing a contracted question with given context assistant: "The question has been contracted. Spawning cove-claim-qs to identify what needs verification." </example>
Use this agent to cross-check claims against independent verification answers and create the final answer. This agent reads claims.md and verifiers/*.md from the atom directory. It compares original claims to independently-obtained answers and creates answer.md with the verification trace. ## Examples: <example> Context: After verifiers have written their responses to files assistant: "I'll spawn cove-verification-maintainer with the atom directory to cross-check claims." </example> <example> Context: Completing factored verification for an atom assistant: "The maintainer will read verifier files and create answer.md with verification results." </example>
External network access
Connects to servers outside your machine
Uses power tools
Uses Bash, Write, or Edit tools
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Advanced reasoning plugin integrating Atom of Thoughts (AoT) and Chain of Verification (CoVe) frameworks.
# Add the marketplace
/plugin marketplace add snowmead/questionably-ultrathink
# Install the plugin
/plugin install questionably-ultrathink@snowmead-marketplace
bunx skills add snowmead/questionably-ultrathink
This installs the skill to your agent's skill directory automatically.
| Command | Description |
|---|---|
/questionably-ultrathink | Full reasoning pipeline (AoT + CoVe) |
/decompose | Break down complex problems into atomic questions |
/verify | Verify factual claims with independent questioning |
The skill automatically activates when you use trigger phrases:
question.md, claims.md, answer.md┌─────────────────────────────────────────────────────────────────┐
│ User Query │
└─────────────────────────────────────────────────────────────────┘
│
▼
┌─────────────────────────────────────────────────────────────────┐
│ Graph Generator │
│ Builds question DAG with dependencies │
│ Creates: atoms/{id}/question.md for each atom │
└─────────────────────────────────────────────────────────────────┘
│
┌──────────────────┼──────────────────┐
▼ ▼ ▼
┌─────────────────┐ ┌─────────────────┐ ┌─────────────────┐
│ Claim Generator │ │ Claim Generator │ │ Claim Generator │
│ (A1) │ │ (A2) │ │ (A3) │
│ [ISOLATED] │ │ [ISOLATED] │ │ [ISOLATED] │
│ Reads: │ │ Reads: │ │ Reads: │
│ question.md │ │ question.md │ │ question.md │
│ Writes: │ │ Writes: │ │ Writes: │
│ claims.md │ │ claims.md │ │ claims.md │
│ verifiers/*.md │ │ verifiers/*.md │ │ verifiers/*.md │
│ (question only)│ │ (question only)│ │ (question only)│
└─────────────────┘ └─────────────────┘ └─────────────────┘
│ │ │
▼ ▼ ▼
┌─────────────────────────────────────────────────────────────────┐
│ Factored Verification │
│ Verifier reads pre-created file containing ONLY the question │
│ Verifier has ZERO context about the original claim │
│ Overwrites: atoms/{id}/verifiers/{N}.md with answer │
└─────────────────────────────────────────────────────────────────┘
│ │ │
▼ ▼ ▼
┌─────────────────┐ ┌─────────────────┐ ┌─────────────────┐
│ Verifier │ │ Verifier │ │ Verifier │
│ "What is X?" │ │ "When was Y?" │ │ "How does Z?" │
│ [NO CONTEXT] │ │ [NO CONTEXT] │ │ [NO CONTEXT] │
└─────────────────┘ └─────────────────┘ └─────────────────┘
│ │ │
└──────────────────┼──────────────────┘
▼
┌─────────────────────────────────────────────────────────────────┐
│ Verification Maintainer │
│ Reads: claims.md + verifiers/*.md │
│ Cross-checks claims vs independent verifier answers │
│ Marks: VERIFIED | REVISED | REFUTED | UNCERTAIN │
│ Creates: answer.md with verification trace │
└─────────────────────────────────────────────────────────────────┘
│
▼
┌─────────────────────────────────────────────────────────────────┐
│ Graph Maintainer │
│ Contracts solved answers INTO dependent questions │
│ Updates question.md: "Given A1=X, A2=Y, what is...?" │
npx claudepluginhub snowmead/questionably-ultrathinkTruth-seeking sparring partner for Claude Code. Challenges claims, decisions, and documents through structured dialectical analysis.
Decomposed reasoning with explicit confidence scoring. Break complex questions into verifiable units, score each independently, and synthesize with uncertainty tracking.
Force Claude to re-validate when you have doubts (!doubt)
Achieve certain comprehension of AI work — /grasp (κατάληψις: a grasping firmly)
AI-augmented Structured Analytic Techniques from US Intelligence Community doctrine — 18 techniques across 6 phases with automated evidence gathering, self-correction, and citation enforcement
Comprehensive UI/UX design plugin for mobile (iOS, Android, React Native) and web applications with design systems, accessibility, and modern patterns