scrying-pool/_bmad-output/implementation-artifacts/review-1-5-group1/README.md

Code Review - Story 1.5 Group 1 (Core Logic)

Generated: 2026-05-22
Story: 1-5-gm-control-ui-scryingpoolstrip-actionpopover-and-av-tile-integration
Review Group: 1 of 4 (Core Logic)
Status: Awaiting parallel reviewer execution

---

REVIEW WORKFLOW

This directory contains prompt files for parallel code review layers. Each layer must be executed in a separate session (ideally with a different LLM model) to ensure independent analysis.

Review Layers

1. 🔴 Blind Hunter (01-blind-hunter-prompt.md)

   • Skill: bmad-review-adversarial-general
   • Input: Diff only (no spec, no context, no project access)
   • Focus: Security issues, bugs, anti-patterns, performance problems
   • Output: Adversarial findings (minimum 5)

2. 🟡 Edge Case Hunter (02-edge-case-hunter-prompt.md)

   • Skill: bmad-review-edge-case-hunter
   • Input: Diff + project read access
   • Focus: Edge cases, boundary conditions, null handling, race conditions
   • Output: Edge case findings (minimum 5)

3. 🟢 Acceptance Auditor (03-acceptance-auditor-prompt.md)

   • Input: Diff + spec file + context docs + project read access
   • Focus: AC compliance, spec violations, missing implementation
   • Output: Acceptance audit findings

---

FILES IN THIS GROUP

New Files (4)

• src/core/ScryingPoolController.js (181 lines)
• src/core/VisibilityManager.js (104 lines)
• tests/unit/core/ScryingPoolController.test.js (277 lines)
• tests/unit/core/VisibilityManager.test.js (218 lines)

Diff Size

• Total: 804 lines (under 3000 line threshold ✓)

---

INSTRUCTIONS FOR REVIEWERS

For Each Layer:

1. Open the prompt file for that layer
2. Read the entire file carefully
3. Execute the review as described in the role section
4. Return findings in the specified format
5. Save findings to a separate file (e.g., findings-blind-hunter.md)

After All Layers Complete:

1. Collect all findings files
2. Return to this workflow at Step 3 (Triage)
3. Paste all findings for consolidated triage

---

REMINDER

• All layers run in parallel — do not wait for one to finish before starting another
• Each layer must use a different session to maintain independence
• All findings are consolidated in Step 3 (Triage) before final presentation

---

NEXT STEPS

After generating findings from all three layers:

1. Proceed to step-03-triage.md in the bmad-code-review workflow
2. Triage and deduplicate findings
3. Present consolidated results