boundary-enumeration

$npx mdskill add yogsoth-ai/de-anthropocentric-research-engine/boundary-enumeration

Systematically tests parameter boundaries to identify claim validity limits

  • Solves the problem of determining valid input ranges for AI artifacts
  • Uses subagents for mapping, generating values, detecting breakpoints, and constructing envelopes
  • Analyzes nominal ranges and generates test values at min, max, and beyond
  • Delivers a structured validity envelope showing where claims hold or break
SKILL.md
.github/skills/boundary-enumerationView on GitHub ↗
---
name: boundary-enumeration
description: "Systematic Boundary Value Analysis: identify parameter boundaries, test at and beyond limits, detect breakpoints."
type: strategy
used-by: [adversarial-stress-testing]
---

# Boundary Enumeration

## Tactics

- boundary-probing

## Method

1. Identify all parameter dimensions of the artifact
2. For each dimension, determine nominal range and boundaries
3. Generate test values at boundaries (min, max, min-1, max+1)
4. Test claim validity at each boundary value
5. Record breakpoints where claim fails
6. Synthesize into validity envelope

## Budget

| Size | Parameter dimensions | Values per dimension | Total probes |
|---|---|---|---|
| S | 3 | 4 | 12 |
| M | 6 | 6 | 36 |
| L | 10 | 8 | 80 |

## Orchestration

1. Dispatch `parameter-space-mapping` to identify dimensions
2. Dispatch `extreme-value-generation` for each dimension
3. Dispatch `breakpoint-detection` at each extreme
4. Dispatch `validity-envelope-construction` to synthesize

## Subagents

- parameter-space-mapping
- extreme-value-generation
- breakpoint-detection
- validity-envelope-construction
More from yogsoth-ai/de-anthropocentric-research-engine