Indian momentum: a replication plan

Using the IIMA Indian Fama–French–momentum library (survivorship-bias-adjusted monthly files): (a) Specify, as code-level rules, how you would test whether Indian WML returns are predictable from the prior 24-month market return and from trailing 6-month WML volatility (the Daniel–Moskowitz "panic state" variables). State the regression, the timing of each variable, and the standard errors. (b) Which months would you inspect as candidate crash episodes, and what pattern would you expect from Proposition 4.7? (c) The IIMA page reports a long-run WML return of 10.65% a year but −5.53% for 2025 up to the December release. Using 4.5, how much information does a single calendar year carry about whether momentum "stopped working"?