AgentSkillsCN

local-stochastic-vol-modeling

适用于校准、曲面动态以及衍生品风险分析的局部与随机波动率建模工作流。当任务涉及为定价与对冲拟合局部波动率/随机波动率模型时，可选用此类工作流。

SKILL.md
--- frontmatter
name: local-stochastic-vol-modeling
description: "Local and stochastic volatility modeling workflows for calibration, surface dynamics, and derivative risk analytics. use when tasks involve fitting local-vol/stochastic-vol models for pricing and hedging."

Local Stochastic Vol Modeling

objective

Calibrate and validate local/stochastic vol models with surface-consistency and hedging diagnostics.

workflow

  1. •define calibration instruments, objective function, and constraints.
  2. •fit local-vol or stochastic-vol parameters with stable optimization.
  3. •validate no-arbitrage surface properties and calibration residuals.
  4. •test hedge performance under dynamic surface scenarios.
  5. •promote only when model risk metrics are within policy limits.

required diagnostics

  • •calibration residual heatmap across strike and tenor.
  • •parameter stability and identifiability diagnostics.
  • •surface smoothness and static-arbitrage checks.
  • •hedge error under scenario-based surface moves.
  • •out-of-sample pricing error across re-calibration windows.

risk controls

  • •enforce parameter bounds and convergence safeguards.
  • •enforce challenger model comparison before release.
  • •enforce fallback model on calibration instability.

outputs

  • •run python scripts/local_stochastic_vol_modeling_diagnostics.py input.csv --output diagnostics.json and keep the json artifact.
  • •write an implementation memo using references/local-stochastic-vol-modeling-playbook.md with assumptions, tests, limits, and rollout plan.

resources

  • •use scripts/local_stochastic_vol_modeling_diagnostics.py for deterministic diagnostics.
  • •use references/local-stochastic-vol-modeling-playbook.md for the domain checklist and delivery structure.