Skip to main content
Glama

scene_add_creature_skill

Destructive

Defines a complex creature's skill by binding an animation to a client graph that triggers at chosen moments, enabling attacks, projectiles, effects, and target-facing behavior in a save copy.

Instructions

Define a skill of a complex creature (造物技能) in a copy of a save: an animation of that creature and a client graph that runs at a chosen moment of it. A complex creature (an asset of the group "怪物(AI版)": scene_search_assets {group:"怪物(AI版)"}) does nothing by itself; what it can do beyond walking is its skills: swing and hit, shoot, dig, hammer at a wall, look around. creature is the asset id of the creature the skill is for (a skill plays the animations of one model and can only be given to that creature). animation is one of its animations, by id or by name ("近战攻击_01"): creature_animations {creature} lists them with their lengths; animations gives several, played one after another (a wind-up, a loop, a recovery). graphId is a graph made with scene_add_graph {kind:"creatureSkill"}; it runs graphAtMs milliseconds into the skill (the moment the weapon lands, the arrow leaves): there it triggers a hit box, fires a projectile or tells a server graph (f.notifyServerNodeGraph, heard as whenSkillNodeIsCalled by a server graph on the creature's own prefab). graphAtMs may be a list of moments, and the graph then runs at each of them: the three strokes of a triple stroke ([350, 900, 1700] for 单手武器_三连击) each trigger its hit box (seen in play; the graph cannot tell which of the moments it is). faceTarget (自动面向目标) turns the creature to its target while the skill plays; interrupts (可打断其他技能) lets the skill cut off another one. effect shows an effect on an attachment point of the creature for part of the animation: the thing the creature holds only while it does this (a pickaxe on GI_WeaponR while it digs, the knife in its hand until it is thrown; what is carried all the time is mounted by a server graph instead, f.mountLoopingSpecialEffect). animationPortion plays only the first part of the animation (0.62 leaves out the two seconds 盗宝团·斥候 just stands at the end of its throw). creature_animations says for the animations that were looked at in play which moment counts and what belongs in the hand; the result repeats it (hint) when graphAtMs is far from that moment. Give the skill to a creature with scene_set_creature_logic, which also sets its cooldown. Writes a new .gil in the workspace; the official save is not touched. A skill of one animation is byte for byte the skill of the official sample 复杂造物; several animations and the two switches are written as the editor saved them on 2026-10-11.

Input Schema

NameRequiredDescriptionDefault
nameYes
effectNo
graphIdYes
mapPathNo
creatureYes
animationNo
graphAtMsNo
animationsNo
faceTargetNo
interruptsNo
outputNameNo
animationPortionNo

Schema Changelog

Changes observed during successful MCP inspections.

  1. First observedv0.1.0

TDQS

A4.7/5.0
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations declare destructiveHint=true, but the description clarifies the actual blast radius: 'Writes a new .gil in the workspace; the official save is not touched,' which changes how an agent should weigh the destructive hint. It also discloses that the graph cannot tell which of several graphAtMs moments it is running at, and that the result repeats a hint when graphAtMs is far from the moment observed in play. It stops short of covering failure modes or mapPath/outputName effects.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The purpose is front-loaded and nearly every parenthetical earns its place by disambiguating a Chinese asset name or giving a concrete value. However, the sentences are extremely long and the closing note about byte-for-byte compatibility with the 2026-10-11 editor save is trivia that slows a reader without changing a call decision.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a 12-parameter mutation tool with a nested object, no output schema and zero schema annotations, the description supplies the graph semantics, the animation-selection model, the effect lifecycle, the downstream consumer (scene_set_creature_logic for cooldown) and the write target. An agent could call this correctly without any other source.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters5/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

With 0% schema description coverage, the description carries the full burden and does so: creature must match the animation's model, animation accepts id or name, animations is a played-in-sequence list, graphAtMs may be a scalar or up to eight moments with a concrete triple-stroke example, effect takes an attachment point like GI_WeaponR, and animationPortion=0.62 trims the trailing idle. Only mapPath and outputName go unexplained.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb and resource: 'Define a skill of a complex creature ... in a copy of a save: an animation of that creature and a client graph that runs at a chosen moment of it.' It immediately distinguishes itself from the sibling scene_add_skill by scoping to complex creatures and from scene_add_graph by naming graphId as an input to this tool.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

It routes the agent explicitly: find the creature with scene_search_assets {group:"怪物(AI版)"}, list animations with creature_animations {creature}, build the graph with scene_add_graph {kind:"creatureSkill"}, then attach the finished skill via scene_set_creature_logic. It also draws a real exclusion line – items held permanently must be mounted by a server graph (f.mountLoopingSpecialEffect) rather than passed as effect.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Deploy Server

Other Tools