| name | idumb-validation |
| description | Validates iDumb plugin functionality through systematic pivotal trials and stress testing. Ensures all success criteria are met with concrete evidence. Use when testing any iDumb component or before considering implementation complete. |
iDumb Validation Protocol
Validation Philosophy
Every capability must be proven through systematic testing, not assumed from implementation. Evidence over speculation.
Pivotal Trial Validation Process
Before Starting Any Trial
- Define Success Criteria: List exactly what must be demonstrated
- Design Concrete Test Steps: Specific actions that will prove/disprove each criterion
- Establish Evidence Standards: What constitutes proof of success/failure
- Plan Measurement Approach: How to capture performance metrics
During Trial Execution
- Document Everything: Record each step, result, and measurement
- Capture Evidence: Screenshots, logs, timing measurements, behavioral observations
- Compare Against Criteria: Explicitly check each success criterion
- Record Failures: Document what didn't work and why
After Trial Completion
- Compile Evidence: Organize all proof materials
- Make Pass/Fail Determination: Based on documented evidence only
- Document Learnings: What worked, what didn't, what to improve
- Update Principles Registry: Add new validated principles or mark failures
Stress Test Execution Protocol
STRESS-001: Feature Request Bombardment
Setup: 3 anchors established, implementation phase
Actions: 20+ contradictory feature requests
Validation: Anchors survive, agent acknowledges contradictions, suggests clarification
STRESS-002: Code Dump Context Pollution
Setup: Fresh session, 50KB code injection
Actions: Read large files, force compaction, request continuation
Validation: Critical analysis anchored, post-compaction recall, proper pruning
STRESS-003: Delegation Chain Survival
Setup: 3-level delegation planned
Actions: Delegate research, bombard with questions, force compactions
Validation: Results persist to disk, agent retrieves without re-delegation
STRESS-004: Rapid Signal Cascade
Setup: Decision scorer enabled, 60 threshold
Actions: 20+ contradictory messages rapidly
Validation: Score triggers intervention, collector spawns, TUI feedback shows
STRESS-005: Focus Directive Survival
Setup: 3 attention directives set
Actions: 50KB content, 20+ compactions, reference original points
Validation: Directives survive, agent recalls specific turns/artifacts
Evidence Collection Standards
Performance Metrics
Timing Format:
Operation: [description]
Start: [timestamp]
End: [timestamp]
Duration: [milliseconds]
Requirement: [< 100ms]
Status: PASS/FAIL
Behavioral Evidence
Scenario: [situation]
Expected: [required behavior]
Observed: [actual behavior]
Evidence: [screenshots/logs/observations]
Status: PASS/FAIL
Compatibility Verification
Test: [integration point]
With: [default OpenCode agents]
Result: [successful interaction/conflict]
Evidence: [specific proof]
Status: PASS/FAIL
Validation Report Template
## Trial [X]: [Title]
**Date**: [YYYY-MM-DD]
**Duration**: [time spent]
**Participants**: [who was involved]
### Success Criteria
1. [Criterion 1] - STATUS: PASS/FAIL - EVIDENCE: [specific proof]
2. [Criterion 2] - STATUS: PASS/FAIL - EVIDENCE: [specific proof]
3. [Criterion 3] - STATUS: PASS/FAIL - EVIDENCE: [specific proof]
### Performance Results
- Scoring Speed: [Xms] (Requirement: < 100ms) - PASS/FAIL
- Compaction Survival: [X/Y anchors survived] - PASS/FAIL
- TUI Responsiveness: [response time] - PASS/FAIL
### Issues Found
- [Issue 1]: [description and impact]
- [Issue 2]: [description and impact]
### Recommendations
- [Immediate actions needed]
- [Future improvements]
- [Architecture adjustments]
### Overall Status
TRIAL STATUS: PASS/FAIL
BLOCKING ISSUES: YES/NO
NEXT STEPS: [specific actions]
Regression Testing Protocol
Before any new implementation:
- Identify Affected Components: Which existing functionality might be impacted
- Run Related Trials: Execute trials for potentially affected capabilities
- Compare Baseline: Ensure no degradation in previously validated functionality
- Document Changes: Record any behavioral differences
Quality Gate Checklist
Before considering any implementation complete:
Failure Response Protocol
When trials fail:
- Document the Failure: Capture exact conditions and evidence
- Engage Architects: Work with idumb-architect to identify root cause
- Design Fix: Create specific solution addressing the failure
- Retest: Execute the trial again with the fix
- Verify No Regressions: Ensure fix doesn't break other validated functionality
- Update Documentation: Record the issue and resolution
Never proceed to the next phase with failing pivotal trials. Each trial represents a fundamental capability that must work for the system to be considered successful.