Morph of PKG

Morph of PKG 1.0.6 — morph walk speeds are now the creature's REAL speed, plus the first measured speed table for modded creatures

Fixed: morph walk speed was built from the wrong multiplier. A morph's speed is derived from how fast the creature actually walks: its base movement speed multiplied by the multiplier the creature's own AI uses. The previous pass read the casual "wander" multiplier — the slow stroll a mob uses when it has nothing to do — instead of the multiplier it uses when it actually wants to get somewhere: chasing you, attacking you, or running away. Those are different numbers, and the second one is the speed a player sees. What that meant in game. Several morphs were quietly too slow. The clearest case is the cat: its wander multiplier is 0.8, but its panic multiplier is 1.5, so a cat morph sat at ~72% of a walking player when a real cat runs faster than that. Cats now move at ~135%, which is how fast a cat actually panics. Same correction for ocelot (120%), spider and cave spider (90%), ravager (90%), pillager and vindicator (123%), llama and trader llama (38% / 61%) and the horse family (63%). Cows read 80% while panicking, pig and sheep 78% / 66%, the skeleton family 75%. Nothing was pushed above the single ceiling. Cat with the fast ability = 189%, pig 78% (it was 250% in the very first pass), rabbit 162% (it was 450%). The one hard 2.0× (200%) ceiling still applies to everything, still enforced in one place. Modded creatures now have a measured speed table — a first. Until now every modded morph fell back to a neutral 1.0 multiplier, because no modded creature had actually been measured. The supported mods' jars were read at bytecode level, so two things are now real numbers instead of a guess: the whole dragon family of one supported dragon mod uses a 1.1 walk multiplier, and that mod's fastest land creature — a small stinging wyvern that strikes with its tail — uses 1.4 on a base movement speed of 0.4, i.e. a real speed of 0.224, which lands exactly on the 200% ceiling. Its variant (the leader form) is measured as well. So the creature users expect to be fast is fast as a morph: the rest of that dragon family reads 176%, one heavier species 151%, and one genuinely sluggish species 44% (slower than a player, correctly). How the modded table stays honest. An entry can target one creature or a whole namespace; an exact entry always wins over a namespace entry, and a namespace entry never leaks onto a different but similarly named namespace. A safety guard (3.0) rejects any absurd multiplier from ever entering the table. Measured values only — anything not measured still falls back to neutral, and the same single ceiling applies. Walking animation follows for free. Leg animation is driven by distance travelled, so a faster morph automatically animates faster. No extra multiplier, no mismatched footwork. Deliberately not changed. The ability key on the stinging wyvern runs its own bite/attack through the source mod's own code path; the effect lands correctly, so no ability or animation change was made there. Gates. The speed gate grew 99 → 116 numeric checks (every table value, the chase / flee / attack reason behind it, modded exact-vs-namespace ordering, namespace safety, and an end-to-end ceiling test on the wyvern), 0 violations. Full build: BUILD SUCCESSFUL, 44 tasks and 0 violations across all gates (speed 116, ability 365, member 1510, language 74, layout 149,516, server 86, transform 33). Known limits: no in-game verification was possible in the build environment; evidence is vanilla source lines, jar bytecode inspection, and the automated numeric gates. Everything else about existing morphs is unchanged.