boundary-probing
$
npx mdskill add yogsoth-ai/de-anthropocentric-research-engine/boundary-probing1. Receive claim and parameter context from strategy 2. Dispatch `parameter-space-mapping` to identify all dimensions 3. For each dimension, dispatch `extreme-value-generation`: - Minimum, maximum, zero, negative, overflow - Type boundaries (empty, null, singleton, infinite) 4. Dispatch `breakpoint-detection` at each extreme value 5. Record pass/fail for each probe point 6. Dispatch `validity-envelope-construction` to synthesize boundaries
SKILL.md
.github/skills/boundary-probingView on GitHub ↗
--- name: boundary-probing description: "Map parameter space, generate extreme values, test at boundaries, detect breakpoints, synthesize validity envelope." type: tactic used-by: [adversarial-stress-testing] strategies: [boundary-enumeration, validity-envelope-mapping, critical-case-design] --- # Boundary Probing ## Orchestration Steps 1. Receive claim and parameter context from strategy 2. Dispatch `parameter-space-mapping` to identify all dimensions 3. For each dimension, dispatch `extreme-value-generation`: - Minimum, maximum, zero, negative, overflow - Type boundaries (empty, null, singleton, infinite) 4. Dispatch `breakpoint-detection` at each extreme value 5. Record pass/fail for each probe point 6. Dispatch `validity-envelope-construction` to synthesize boundaries ## Subagents - parameter-space-mapping - extreme-value-generation - breakpoint-detection - validity-envelope-construction ## Termination Conditions - All dimensions probed at all boundary values (complete) - Breakpoint found in critical dimension (early report) - Budget exhausted (report partial envelope) - Dimension count exceeds budget (prioritize by sensitivity)