| name | amaa-hypothesis-verification |
| description | Use when verifying technical claims through Docker-isolated experiments. Trigger with hypothesis test or experiment request. Loaded by ai-maestro-architect-agent-main-agent |
| agent | ai-maestro-architect-agent-main-agent |
| context | fork |
| user-invocable | false |
Hypothesis Verification Skill
Overview
Verify claims through controlled Docker experimentation. Everything is "To Be Verified" (TBV) until you personally test it. Claims from any source require experimental confirmation.
Prerequisites
- Docker installed and running
- Write access to experiment output directories
- Understanding of the claim to be verified
Checklist
Copy this checklist and track your progress:
Instructions
- Identify the claim and mark as TBV
- Set up Docker container for isolated testing
- Design experiment with 3+ approaches (Multiplicity Rule)
- Execute experiments and collect measurements
- Classify: VERIFIED / UNVERIFIED / PARTIALLY VERIFIED
- Document findings in experimentation report
- Clean up containers; archive prototype if valuable
Reference Documents
| Document | Content |
|---|
| docker-experimentation.md | Container setup and templates (Why Docker is Required, Container Structure Template, docker-compose.yml Template, Container Cleanup Procedure) |
| researcher-vs-experimenter.md | The Researcher (What OTHERS say is true), The Experimenter (What I can PROVE is true), The TBV Principle (To Be Verified), Workflow Integration: Researcher → Experimenter |
| experiment-scenarios.md | Case 1: Post-Research Validation, Case 2: Issue Reproduction in Isolation, Case 3: Architectural Bug Investigation, Case 4: New API/Tool Evaluation, Case 5: Fact-Checking Claims (Quick Verification) |
Examples
Claim: "Redis caches responses 10x faster than in-memory dict" (TBV)
-> Docker container with Redis + Python
-> Test 3 approaches: dict, Redis, Redis pooled (1000 iterations)
-> Result: Dict 0.001ms, Redis 0.15ms, Redis pooled 0.08ms
-> Classification: UNVERIFIED (Redis slower for simple cases)
Error Handling
| Error | Solution |
|---|
| Docker not available | Start Docker Desktop or docker service |
| Container cleanup failed | Run docker system prune |
| Experiment inconclusive | Increase sample size, reduce variables |
| Conflicting results | Standardize container configuration |
Output
| Artifact | Location |
|---|
| Experimentation Report | experiments/<claim-name>/REPORT.md |
| Measurement Data | experiments/<claim-name>/data/ |
| Prototype Archive | prototypes/<claim-name>/ (if valuable) |
Resources
See Reference Documents table above.
- amaa-hypothesis-verification-ops - operations and troubleshooting quick-reference companion to this skill (loaded on demand for runtime ops/debugging).