assumption-constraint · diff
v1.0.0 to v1.0.0
9 added, 3 removed. Audit A to A.
---
name: assumption-constraint
- description: "Which assumptions are most fragile? — Vulnerability ranking + impact assessment of experiment assumptions"
+ description: Which assumptions are most fragile? — Vulnerability ranking + impact
+ assessment of experiment assumptions
version: 1.0.0
category: experiment-execution
type: strategy
- used-by: constraint-analysis
sops:
+ - assumption-challenging
+ - resource-quantification
+ tactics:
+ - sensitivity-ranking
+ dependencies:
+ sops:
- assumption-challenging
- resource-quantification
- tactics:
+ tactics:
- sensitivity-ranking
---
# Strategy: Assumption Constraint
## Methodology
Systematic assumption vulnerability analysis:
- **Extraction**: Surface all implicit and explicit assumptions
- **Scoring**: Quantify vulnerability (confidence × evidence / testability)
- **Impact assessment**: Blast radius × recovery cost
- **Prioritization**: Vulnerability × Impact = Priority
- **Validation planning**: Cheapest test that resolves uncertainty
Assumption categories:
| Category | Examples |
|----------|----------|
| Technical | Method convergence, architecture suitability |
| Data | Availability, quality, representativeness |
| Resource | Sufficiency of compute, time, expertise |
| Environmental | Tool stability, API access, policy |
| Theoretical | Effect existence, measurability, magnitude |
## Execution Flow
1. **Challenge Assumptions** → call `assumption-challenging` SOP
- Input: experiment plan, hypothesis
- Output: assumption inventory with validity assessment
2. **Quantify Validation Cost** → call `resource-quantification` SOP
- Input: validation experiments for top assumptions
- Output: cost to validate each assumption
3. **Rank Sensitivity** → invoke `sensitivity-ranking` tactic
- Determine which assumptions are most binding
4. **Report** → synthesize vulnerability assessment
- Top-5 fragile assumptions with validation paths
- Binding assumption constraint identification
## Budget Gate
| Resource | Budget | Notes |
|----------|--------|-------|
| Subagent calls | ≤5 | 2 SOPs + synthesis |
| Iterations | ≤2 | Re-rank if new assumptions surface |
| Output size | ≤3000 tokens | Ranked table + validation plan |