Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Simulate a 5-member expert board deliberation for major decisions. Use when evaluating plans, architecture choices, feature designs, or any decision requiring multi-perspective expert analysis. Triggers: 'board review', 'get expert opinions', 'board meeting', 'director evaluation', 'consensus review'.
.claude/skills/ibrahim-3d-board-of-directors/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | 71% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 44% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 17% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 22% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 56% | 0% |
Simulates a 5-member expert board that deliberates, debates, and reaches consensus on major decisions. Each director brings domain expertise and can challenge other directors' opinions.
| Role | Domain | Evaluates | |------|--------|-----------| | Chief Architect (CA) | Technical | System design, patterns, scalability, tech debt, code quality | | Chief Product Officer (CPO) | Product | User value, market fit, feature priority, scope, usability | | Chief Security Officer (CSO) | Security | Vulnerabilities, compliance, data protection, risk assessment | | Chief Operations Officer (COO) | Execution | Feasibility, timeline, resources, process, deployment | | Chief Experience Officer (CXO) | Experience | UX/UI, accessibility, user journey, design consistency |
Each director reviews the proposal independently:
DISPATCH via Task tool (all 5 in parallel):
- CA: Evaluate technical aspects
- CPO: Evaluate product aspects
- CSO: Evaluate security aspects
- COO: Evaluate operational aspects
- CXO: Evaluate experience aspectsEach director outputs:
json{ "director": "CA", "verdict": "APPROVE" | "CONCERNS" | "REJECT", "score": 1-10, "key_points": ["..."], "concerns": ["..."], "questions_for_board": ["Question for CPO about...", "Challenge to COO on..."] }
Directors respond to each other's questions and challenges:
MESSAGE BUS: conductor/tracks/{track}/.message-bus/board/
1. Post all Phase 1 assessments to board/assessments.json
2. Each director reads others' assessments
3. Directors post rebuttals/responses to board/discussion.jsonl
4. Max 3 rounds of discussionDiscussion message format:
json{ "from": "CA", "to": "CPO", "type": "CHALLENGE" | "AGREE" | "QUESTION" | "CLARIFY", "message": "Regarding your concern about scope...", "changes_my_verdict": true | false }
After discussion, each director casts final vote:
json{ "director": "CA", "final_verdict": "APPROVE" | "REJECT", "confidence": 0.0-1.0, "conditions": ["Must add rate limiting", "Needs load testing"], "dissent_noted": false }
Aggregate votes and produce board decision:
| Scenario | Resolution | |----------|------------| | 5-0 or 4-1 APPROVE | APPROVED — Proceed with any conditions noted | | 3-2 APPROVE | APPROVED WITH REVIEW — Proceed but schedule follow-up | | 3-2 REJECT | REJECTED — Address major concerns first | | 4-1 or 5-0 REJECT | REJECTED — Significant rework needed | | 2-2-1 (tie with abstain) | Chief Architect (CA) casts tiebreaking vote based on technical merit |
After reaching resolution, you MUST persist the decision to files:
mkdir -p conductor/tracks/{trackId}/.message-bus/board/resolution.md with the Board Output Format (below)session-{timestamp}.json:json {"session_id": "...", "verdict": "...", "vote_summary": {...}, "conditions": [...], "timestamp": "..."}
Then return ONLY this concise summary to the orchestrator:
json{"verdict": "APPROVED|REJECTED|ESCALATE", "conditions": ["..."], "vote": "4-1"}
typescriptasync function invokeBoardReview(proposal: string, context: object) { // 1. Initialize board message bus await initBoardMessageBus(trackId); // 2. Phase 1: Parallel assessment const assessments = await Promise.all([ Task({ description: "CA board assessment", prompt: `You are the Chief Architect on the Board of Directors. PROPOSAL: ${proposal} CONTEXT: ${JSON.stringify(context)} Follow the directors/chief-architect.md profile. Output your assessment as JSON.` }), Task({ description: "CPO board assessment", ... }), Task({ description: "CSO board assessment", ... }), Task({ description: "COO board assessment", ... }), Task({ description: "CXO board assessment", ... }) ]); // 3. Phase 2: Discussion rounds await runBoardDiscussion(assessments, maxRounds: 3); // 4. Phase 3: Final vote const votes = await collectFinalVotes(); // 5. Phase 4: Resolution return aggregateBoardDecision(votes); }
markdown## Board of Directors Resolution **Proposal**: [Brief description] **Session**: [timestamp] **Verdict**: APPROVED | APPROVED WITH REVIEW | REJECTED | ESCALATE ### Vote Summary | Director | Vote | Confidence | Key Condition | |----------|------|------------|---------------| | CA | APPROVE | 0.9 | Add caching layer | | CPO | APPROVE | 0.8 | Validate with usability check | | CSO | CONCERNS→APPROVE | 0.7 | Security audit before launch | | COO | APPROVE | 0.85 | Need 2-week buffer | | CXO | APPROVE | 0.95 | Accessibility is solid | **Final: 5-0 APPROVE** ### Conditions for Approval 1. Add caching layer for API responses (CA) 2. Complete security audit before production (CSO) 3. Buffer timeline by 2 weeks (COO) ### Discussion Highlights - CA challenged CPO on scope creep → CPO agreed to defer Phase 2 - CSO raised auth concern → CA proposed token rotation solution - CXO praised accessibility approach, no concerns ### Dissenting Opinions None recorded. --- *Board session complete. Proceed with implementation.*
Each director has specialized evaluation criteria. See:
directors/chief-architect.md — Technical excellencedirectors/chief-product-officer.md — Product valuedirectors/chief-security-officer.md — Security posturedirectors/chief-operations-officer.md — Execution realitydirectors/chief-experience-officer.md — User experienceFor rapid board review without full deliberation:
markdown/board-review [proposal] Returns: Quick assessment from each director (no discussion phase)
For full deliberation:
markdown/board-meeting [proposal] Returns: Full 4-phase deliberation with discussion
The board can be invoked at key checkpoints:
| Checkpoint | Board Involvement | |------------|-------------------| | EVALUATE_PLAN | Full board meeting for major tracks | | EVALUATE_EXECUTION | Quick review for implementation quality | | Pre-Launch | Security + Operations deep dive | | Post-Mortem | All directors analyze what went wrong |
.message-bus/board/
├── session-{timestamp}.json # Session metadata
├── assessments.json # Phase 1 outputs
├── discussion.jsonl # Phase 2 messages
├── votes.json # Phase 3 final votes
└── resolution.md # Phase 4 board decision| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 40,645 | 8,538 | -79% | 1 | 1 | 0% | 6,636 | 2,489 | -62% | 0 | 0 | — |
case-02 | fail→fail | 5,473 | 6,951 | +27% | 1 | 1 | 0% | 339 | 2,560 | +655% | 0 | 0 | — |
case-03 | fail→fail | 10,522 | 7,006 | -33% | 1 | 1 | 0% | 233 | 2,406 | +933% | 0 | 0 | — |
case-04 | fail→pass | 8,663 | 2,125 | -75% | 1 | 1 | 0% | 1,353 | 2,316 | +71% | 0 | 0 | — |
case-05 | pass→pass | 6,721 | 2,042 | -70% | 1 | 1 | 0% | 999 | 2,278 | +128% | 0 | 0 | — |
case-06 | fail→pass | 9,963 | 1,540 | -85% | 1 | 1 | 0% | 1,503 | 2,166 | +44% | 0 | 0 | — |
case-07 | fail→pass | 14,883 | 6,391 | -57% | 1 | 1 | 0% | 2,484 | 2,906 | +17% | 0 | 0 | — |
case-08 | fail→pass | 13,705 | 3,837 | -72% | 1 | 1 | 0% | 2,063 | 2,516 | +22% | 0 | 0 | — |
case-09 | fail→pass | 7,745 | 2,253 | -71% | 1 | 1 | 0% | 1,481 | 2,303 | +56% | 0 | 0 | — |
case-10 | fail→pass | 10,352 | 2,698 | -74% | 1 | 1 | 0% | 1,652 | 2,397 | +45% | 0 | 0 | — |
case-11 | fail→pass | 8,587 | 3,416 | -60% | 1 | 1 | 0% | 1,262 | 2,520 | +100% | 0 | 0 | — |
case-12 | fail→pass | 11,404 | 2,936 | -74% | 1 | 1 | 0% | 624 | 2,460 | +294% | 0 | 0 | — |
case-13 | fail→pass | 19,730 | 3,133 | -84% | 1 | 1 | 0% | 1,069 | 2,523 | +136% | 0 | 0 | — |
case-14 | fail→pass | 6,658 | 1,538 | -77% | 1 | 1 | 0% | 1,008 | 2,134 | +112% | 0 | 0 | — |
case-15 | fail→pass | 6,745 | 1,266 | -81% | 1 | 1 | 0% | 1,081 | 2,038 | +89% | 0 | 0 | — |
case-16 | fail→pass | 7,334 | 1,195 | -84% | 1 | 1 | 0% | 1,145 | 2,056 | +80% | 0 | 0 | — |
case-17 | pass→pass | 8,721 | 1,221 | -86% | 1 | 1 | 0% | 1,314 | 2,091 | +59% | 0 | 0 | — |
case-18 | pass→pass | 5,372 | 1,952 | -64% | 1 | 1 | 0% | 761 | 2,198 | +189% | 0 | 0 | — |
case-19 | fail→pass | 4,830 | 5,780 | +20% | 1 | 1 | 0% | 739 | 3,045 | +312% | 0 | 0 | — |
case-20 | fail→fail | 9,639 | 2,137 | -78% | 1 | 1 | 0% | 1,564 | 2,247 | +44% | 0 | 0 | — |
case-21 | fail→pass | 10,585 | 1,335 | -87% | 1 | 1 | 0% | 1,466 | 2,117 | +44% | 0 | 0 | — |
case-22 | pass→pass | 7,613 | 7,257 | -5% | 1 | 1 | 0% | 1,370 | 3,206 | +134% | 0 | 0 | — |
case-23 | pass→pass | 14,784 | 12,698 | -14% | 1 | 1 | 0% | 2,461 | 3,981 | +62% | 0 | 0 | — |
case-24 | pass→pass | 14,166 | 24,460 | +73% | 1 | 1 | 0% | 2,755 | 6,084 | +121% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 24 cases were attempted, and 19 counted toward the lift figure. The other 5 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +58 percentage points is the difference between those two pass rates over the 19 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.