notque/vexjoy-agent / skills/analysis/assessment
assessment
VerifiedAssessment: read-only inspection, codebase overview, value analysis, health checks, ADR consultation, decision analysis, multi-perspective critique.
references/multi-persona-critique/synthesis-template.md
View on GitHub# Multi-Persona Critique: Synthesis Report Template
Use this template for the final synthesis report output in Phase 5.
## Report Structure
```markdown
# Multi-Persona Critique Report
**Proposals evaluated:** [N]
**Personas consulted:** The Logician, The Pragmatic Builder, The Systems Purist, The End User Advocate, The Skeptical Philosopher
**Methodology:** 5 independent parallel critiques with consensus synthesis
---
## 1. Consensus Matrix
| # | Proposal | Logician | Builder | Purist | User Advocate | Philosopher | Score | Pattern |
|---|----------|----------|---------|--------|---------------|-------------|-------|---------|
| 1 | [Title] | STRONG | STRONG | PROMISING | STRONG | PROMISING | 13 | CONSENSUS |
| 2 | [Title] | PROMISING| REJECT | STRONG | WEAK | REJECT | 6 | CONTESTED |
| 3 | [Title] | WEAK | PROMISING| WEAK | STRONG | WEAK | 7 | OUTLIER |
**Scoring:** STRONG=3, PROMISING=2, WEAK=1, REJECT=0. Max possible: 15.
**Pattern definitions:**
- CONSENSUS: 4+ personas agree within one tier
- CONTESTED: 2-3 way split across tiers
- OUTLIER: 1 persona disagrees with 4
---
## 2. The Features You Should Build
*Proposals scoring 12-15 — survived scrutiny from all 5 lenses.*
### [Proposal Title] (Score: N/15, Pattern: CONSENSUS)
**Why all 5 lenses approve:**
- Logician: [key justification]
- Builder: [key justification]
- Purist: [key justification]
- User Advocate: [key justification]
- Philosopher: [key justification]
**Conditions noted (even in agreement):**
- [Any caveats from personas who rated PROMISING rather than STRONG]
[Repeat for each proposal scoring 12+]
*If no proposals score 12+: "No proposals achieved near-unanimous STRONG consensus. This is informative — the proposal set may need refinement before commitment."*
---
## 3. Worth Investigating
*Proposals scoring 8-11 — promising core with specific conditions to address.*
### [Proposal Title] (Score: N/15, Pattern: [CONSENSUS/CONTESTED])
**What makes it promising:**
- [Persona]: [strongest argument for]
- [Persona]: [second strongest argument]
**What would make it STRONG:**
- [Persona who rated lower]: [specific condition that would upgrade their rating]
- [Persona who rated lower]: [specific condition]
**Decision guidance:** [One sentence on what the user should investigate before committing]
[Repeat for each proposal scoring 8-11]
---
## 4. Interesting Disagreements
*Proposals where personas split — the disagreements reveal real tradeoffs.*
### [Proposal Title] (Score: N/15, Pattern: CONTESTED)
**The split:**
| For (rated STRONG/PROMISING) | Against (rated WEAK/REJECT) |
|------------------------------|---------------------------|
| [Persona]: [argument] | [Persona]: [counter-argument] |
| [Persona]: [argument] | [Persona]: [counter-argument] |
**Nature of disagreement:** [Values disagreement (what matters) / Factual disagreement (what is true) / Perspective disagreement (different lens sees different things)]
**What each side sees that the other does not:**
- The [For] side sees: [insight]
- The [Against] side sees: [insight]
**Why this matters:** [1-2 sentences on what the tension reveals about the proposal]
[Repeat for each contested proposal]
---
## 5. Shelve
*Proposals scoring 0-7 — did not survive this gauntlet.*
| Proposal | Score | Most Critical Persona | Core Reason |
|----------|-------|----------------------|-------------|
| [Title] | N/15 | [Persona] | [1-sentence reason] |
**Important:** "Shelve" does not mean "bad idea." It means the idea did not survive critique from 5 independent philosophical lenses in its current form. Reformulation or context change could alter the outcome.
---
## 6. Cross-Cutting Insights
*Meta-observations that cut across multiple proposals.*
### Themes Surfaced
- [Theme 1]: [Which proposals share this theme, and what it reveals]
- [Theme 2]: [Which proposals share this theme, and what it reveals]
### Blind Spots in the Proposal Set
- [What is missing?]: [Gap that no proposal addresses but the personas collectively noticed]
### Tensions Between Proposals
- [Proposal X vs Proposal Y]: [How building one undermines or complicates the other]
---
## 7. The Deepest Insight
[What did the critics collectively reveal that was not obvious from any single perspective? This is the paragraph that makes the entire exercise worth doing. It should synthesize across all personas and all proposals to surface something the proposer could not have seen alone.]
---
## Caveats
- All 5 personas are generated by the same underlying model, which means they share certain blind spots (training data biases, inability to verify empirical claims, limited knowledge of the specific organizational context)
- Consensus among model-generated personas is weaker evidence than consensus among actual humans with diverse lived experience
- This analysis evaluates reasoning quality and identifies tensions — it does not replace domain expertise, user research, or empirical testing
```
## Section Population Rules
When filling in the template:
1. **Consensus Matrix**: Include ALL proposals and ALL personas. No omissions. Calculate scores arithmetically.
2. **Features You Should Build**: Only proposals scoring 12+. If none qualify, say so explicitly — an empty section is honest; a lowered bar is not.
3. **Worth Investigating**: Score 8-11 only. Focus on the specific conditions that would upgrade the rating. Be concrete ("the Builder needs a migration path") not vague ("needs more thought").
4. **Interesting Disagreements**: The most valuable section. Present both sides with equal respect and identify the nature of the disagreement (values vs facts vs perspective). Leave resolution to a later synthesis section unless the personas already reached a shared conclusion.
5. **Shelve**: Brief and respectful. One sentence per proposal. No mockery.
6. **Cross-Cutting Insights**: Look for patterns ACROSS proposals, not within them. Themes, gaps, and tensions.
7. **The Deepest Insight**: One paragraph. This is the synthesis of synthesis — what emerges from the full exercise that no single persona or single proposal could have revealed. If nothing emerges, say "No cross-cutting insight beyond the individual critiques" rather than fabricating profundity.
## Tone Guidelines
| Instead of | Write |
|------------|-------|
| "This idea is terrible" | "This proposal scored 2/15. The Logician and Philosopher both identified [specific flaw]." |
| "Obviously the best choice" | "This scored 14/15 with consensus across all 5 lenses, though the Builder noted [caveat]." |
| "The personas couldn't agree" | "This is contested (score 8/15) with a values-level disagreement between [personas]." |
| "Everyone agrees" | "Consensus at STRONG (score 15/15). Note: shared-model limitation applies." |
**Preserve from persona outputs:** Specific reasoning, evaluation criteria references, conditional statements.
**Do not alter:** Individual persona ratings or justifications. The synthesis adds analysis on top; it never edits what the personas said.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| receiving-code-reviewobra/superpowersUse when receiving code review feedback, before implementing suggestions, especially if feedback seems unclear or technically questionable - requires technical rigor and verification, not performative agreement or blind implementation | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| requesting-code-reviewobra/superpowersUse when completing tasks, implementing major features, or before merging to verify work meets requirements | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| using-git-worktreesobra/superpowersUse when starting feature work that needs isolation from current workspace or before executing implementation plans - ensures an isolated workspace exists via native tools or git worktree fallback | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| verification-before-completionobra/superpowersUse when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions always | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |