| name | minecraft-custom-mod-builder |
| description | Minecraft Custom Mod Builder: Generate and verify structured Minecraft Bedrock, Bedrock skin pack, Fabric, and NeoForge projects. Use when an agent needs minecraft custom mod builder, create runtime verified bedrock add ons and skin packs; build and verify fabric and neoforge jars; generate typed items, blocks, entities, commands, create mod project, target platform, mod id through AgentPMT-hosted remote tool calls. Discovery terms: minecraft custom mod builder. |
| version | 1.0.8 |
| homepage | https://www.agentpmt.com/marketplace/minecraft-custom-mod-builder |
| compatibility | Agent instructions for AgentPMT-hosted remote tool calls. Follow this skill body for supported account, wallet, and setup routes. No local command runtime is declared. |
| metadata | {"author":"agentpmt","openclaw":{"homepage":"https://www.agentpmt.com/marketplace/minecraft-custom-mod-builder"}} |
Minecraft Custom Mod Builder
Freshness
Last updated: 2026-07-14.
If the current date is more than 7 days after the last updated date, reinstall this skill from skills.sh or ClawHub before relying on endpoints, schemas, setup steps, or examples.
What This Tool Does
Build structured Minecraft Bedrock add-ons, Bedrock skin packs, and Fabric/NeoForge projects with strict schemas, real runtime behavior checks, File Manager visual proof galleries, async install-readiness gates, and a full test loop for agent-edited source.
Product Instructions
Minecraft Mod Builder
Generate deterministic Minecraft Bedrock, Fabric, and NeoForge mod artifacts from structured specifications. This tool does not call an LLM, does not accept prompts, and does not require user credentials.
Two Rules That Will Save You Many Retries
Rule 1 — One ItemSpec equals one in-game item
A single ItemSpec produces one in-game item. The fields damage, durability, max_stack_size, nutrition, tool, armor, food, block_placer, entity_placer, enchantable, repairable, digger, cooldown, glint, rarity, dyeable, and component_overrides all attach to the same item. A weapon does not need a separate entity or a separate "damage item." A sword does not need a partner "enchantment item." You ship one item; this tool wires up every relevant Minecraft component on that item from the fields above.
Do not split one weapon idea across many items[] entries. Do not duplicate a weapon and call one copy "damage" and another "durability." That is the most common source of bloated, broken submissions.
Rule 2 — Every visible asset requires a texture
Items, blocks, entities, particles, named texture assets, machines, storage, and skin-pack skins must each carry a texture with one of these three sources:
source_base64 — base64-encoded PNG or JPEG bytes (max 1024×1024, max 5 MB).source_file_id — a File Manager file_id for a user-uploaded image.color_hex — an explicit #RRGGBB color. Only item_kind in {tool, weapon} (with tool_type) and the dedicated texture_kind in {block, entity} paths produce a recognizable sprite (sword/pickaxe/axe/shovel/hoe; shaded block face; model-compatible cuboid atlas). For every other kind — generic, food, fuel, projectile, block_placer, entity_placer, armor, particles, machines, storage — color_hex renders only as a flat colored 16×16 square. That is appropriate for raw materials (ingots, dust, gems) but looks like a missing texture for armor/food/machines. Use source_base64 or source_file_id whenever the item's identity is visual (armor, food, machines, storage, projectiles). Tool/weapon color_hex requires tool_type; otherwise rejected.
A spec that omits the texture binding is rejected at validation with MINECRAFT_VISIBLE_ASSET_TEXTURE_REQUIRED. There is no procedural blob fallback. If you want the user to see something, you supply the texture.
Java armor inventory icons must ship a real PNG. item_kind="armor" rejects color_hex-only textures (MINECRAFT_ARMOR_TEXTURE_REQUIRES_IMAGE) because that produces only a flat inventory square. Supply source_base64 or source_file_id for every helmet/chestplate/leggings/boots. Declarative Java armor uses the selected vanilla iron/gold/diamond/netherite equipment asset while worn; the supplied image is the inventory icon, not a wearable UV atlas. A custom material or custom worn texture must use the agent-edited source workflow and pass client/runtime verification. Declarative Bedrock armor is rejected because a wearable item without a verified attachable would be invisible on the wearer.
Authoring textures with the Icon Generator + pixelize_to
The canonical way to author textures for items, blocks, entities, and particles is the Icon Generator (/product-icon-generator). Render at the exact target grid in pixel_art_mode=true — 16×16 for items and blocks, 64×64 for entities and skin packs — and pipe the returned file_id straight into this texture's source_file_id. Detailed recipes live in agent_instructions/minecraft_texture_agent.md.
When the texture originates outside the Icon Generator (AI image creator, user upload, third-party source), use the ingest-side safety net by setting texture.pixelize_to and optionally texture.palette_colors. Both fields are opt-in (None by default) so existing bindings remain byte-identical. Example: {"texture": {"source_file_id": "abc...", "pixelize_to": 16, "palette_colors": 8}} nearest-neighbor downscales the input to 16×16 and quantizes it to an 8-color palette with dithering disabled, producing crisp pixel-art output regardless of how the source was made. Failures surface as MINECRAFT_TEXTURE_PIXELIZE_FAILED.
Recommended defaults unless you have a specific reason to deviate: items/blocks → pixelize_to: 16, palette_colors: 8; entities/skins → pixelize_to: 64, palette_colors: 16; particles → pixelize_to: 16, palette_colors: 4. The color_hex binding path is unaffected — it already produces crisp procedural sprites.
Request Shape Rules
Prefer the canonical field names below. The backend accepts the listed aliases when they are exact equivalents, but canonical names make the generated schema and validation errors easier to follow.
Verified async builds and install readiness
Use start_build_job for anything the user should install. It validates the request, queues the build, and returns immediately with task_id and status="queued". Poll with get_build_job using the same task_id until status is completed or failed.
start_build_job defaults to verification_level="behavior". The completed job result includes:
runtime_verification — boot/behavior/render/client checks or the exact reason verification could not run.visual_proof — File Manager–hosted PNG frames (texture enlargements, entity composites, particle sheets, and Java in-game client screenshots when the client lane runs). Each frame has signed_url, file_id, caption, and machine checks. Video clips are not part of this surface yet.quality_gate — passed, failed, blocked, or unverified. Visual fidelity/frame failures use MINECRAFT_VISUAL_FIDELITY_INSUFFICIENT or MINECRAFT_VISUAL_PROOF_FAILED.ready_for_install — true only when the quality gate passed (behavior and required visual proof).verification_not_run — true when the requested runtime checks did not execute.request_classification — supported_declarative, needs_agent_code, or platform_constrained.
Agent visual review is mandatory. Before telling a user a mod is finished, open every visual_proof.frames[].signed_url (or use the File Manager tool with each frame file_id under the same budget) and confirm the art matches the request. The same frames are also listed in artifacts[] with label=visual_proof_frame. Do not claim success from behavior checks alone. If frames show flat colored squares, missing identity art, wrong subjects, or unusable scale, fix textures (source_base64 / source_file_id, not only color_hex for armor, food, entities, particles, machines, or skins) or edit source and re-run start_build_job.
Treat ready_for_install=false, quality_gate.status!="passed", verification_not_run=true, or visual_proof.status in {failed,partial} on a behavior job as "fix the spec/art/source and rerun." Do not present those artifacts as working mods even if the task status is completed; completed only means the job produced inspectable output.
A generated behavior job must declare at least one concrete feature, advanced resource, or skin. Use create_mod_project or verification_level="off" for an intentionally empty scaffold; an empty project cannot be certified as a functional mod.
Use list_build_jobs with limit (1 to 100) to inspect recent build jobs for the current budget.
create_mod_project remains synchronous and unverified. It accepts only verification_level="off" and returns ready_for_install=false. Use it for fast source/debug output, not final user delivery.
Use verification_level="boot_smoke" only when the user explicitly wants a faster runtime-load check instead of behavior checks. Use verification_level="off" only for debug/source iteration and tell the user the result is unverified.
Agent-edited source testing
For behavior the declarative schema cannot faithfully express, generate source first, edit the code, upload a zip, then call start_build_job with:
{
"action": "start_build_job",
"target_platform": "fabric",
"mod_id": "edited_mod",
"authoring_mode": "agent_code",
"source_archive_file_id": "file_...",
"source_test_mode": "test_only",
"verification_contract": {
"schema_version": "1",
"target_platform": "fabric",
"mod_id": "edited_mod",
"user_intent_summary": "Using the edited mod marks the player as verified.",
"expected_checks": [
{
"check_id": "boot",
"description": "The edited mod boots without registration errors."
source_test_mode defaults to test_only; use that mode while iterating. Set source_test_mode="test_and_package" only when the runtime report passes; install artifacts are returned only when ready_for_install=true. Do not put verifier files under .agentpmt/verification/** into shipped artifacts.
For verification_level="behavior", the contract must include at least one behavior, render, or client check. A contract containing only boot and/or static checks is rejected because loading and file presence do not prove requested behavior. Use verification_level="boot_smoke" when load-only verification is the actual goal; boot-smoke output is never install-ready.
authoring_mode is deterministic: auto and declarative validate the typed feature schema; agent_code requires source_archive_file_id on start_build_job. The service never guesses this route from keywords in a prompt.
Every edited-source check must be machine-observable. Contract assertion kinds are grouped by surface:
check_kind | Allowed assertion_kind values |
|---|
boot | boot_loaded |
behavior | command_succeeds, scoreboard_value, scoreboard_delta, tag_present, tag_absent, effect_present, block_state, entity_present, entity_spawned, item_present, item_granted, item_dropped, teleport_delta, time_window, weather_state, machine_output |
static | file_exists |
render | render_coverage |
client | client_option_state, message_observed, title_observed, screenshot_marker |
Use setup_commands for bounded Minecraft-world setup and trigger_command for the command that causes the behavior. scoreboard_value and scoreboard_delta targets are objective or objective:holder; item, entity, effect, and block assertions use namespaced identifiers; teleport_delta.expected_value is an x,y,z offset; time_window.target uses the documented time keywords; weather_state.target is clear, rain, or thunder; screenshot_marker.target is #RRGGBB and its optional expected_value is a minimum pixel count. message_observed and title_observed compare exact text from the actual Java client GUI. screenshot_marker inspects the framebuffer. Self-reported log/client markers are not accepted as proof. The nonce-bound AGENTPMT_RUNTIME marker family is harness-private and rejected in uploaded source.
When typed postconditions cannot trigger the edited code directly, include the applicable verifier overlay in the source zip. Java server checks use .agentpmt/verification/java/AgentpmtCustomVerifier.java with public static void verify(String checkId, ServerLevel serverLevel, ServerPlayer player). Java client checks may add .agentpmt/verification/java/AgentpmtCustomClientVerifier.java with public static void verify(String checkId, Minecraft client). Bedrock uses .agentpmt/verification/bedrock/custom_verifier.js with export function verifyCustomCheck(checkId, context). A Java archive may include both server and client verifier files when its contract has both surfaces. The harness always evaluates its built-in postcondition first, then runs a supplied custom verifier before recording success; agent-supplied verifier code is never the sole proof. The verifier must throw on failure and cannot replace a generic postcondition. Verifier overlays and .agentpmt/runtime/** are reserved, throwaway paths and never ship in artifacts.
For client_option_state, set target to render_distance, simulation_distance, entity_distance_scaling, gamma, graphics_mode, cloud_status, or particles, and set expected_value to the exact requested value. Use message_observed/title_observed for text and screenshot_marker with a #RRGGBB target for arbitrary visual overlays or particles; do not ask verifier code to certify its own output.
Uploaded Java build scripts, wrappers, buildSrc, alternate Gradle hooks, generated output/cache directories (build, .gradle, out, target), and precompiled Java/native binaries are not trusted. The service rejects generated outputs and precompiled .class/.jar/native-library files, replaces executable Gradle configuration with the pinned loader template, passes an explicit allowlisted child environment, and gives each job a disposable Gradle home backed by read-only shared caches. Submit editable source and ordinary resource assets only. Agent-edited source executes only in a private disposable Cloud Run worker with a dedicated no-role service account and one verification request per process. If that worker is unavailable, verification fails closed and the artifact is not install-ready.
advanced_resources is for non-executable data/resource files and Bedrock .mcfunction files. It rejects Java source and behavior_pack/scripts/*.js. To change executable Fabric, NeoForge, or Bedrock Script API code, edit the generated project and submit the complete zip with authoring_mode="agent_code"; do not try to inject executable code through a declarative request.
Bedrock custom namespace
For Bedrock add-ons, generated custom in-game identifiers use the Creator Tools namespace creator_project:<id>. Keep using mod_id for the request, artifact naming, file paths, and Java platforms. In Bedrock specs, short ids ("ruby") and legacy <mod_id>:<id> refs still resolve to the generated custom feature; the emitted pack rewrites them to creator_project:<id>. Use minecraft:<id> for vanilla content and other explicit namespaces for external content.
Creative world mechanics and fidelity
Prank, cursed, chaotic, fake-glitch, admin, debug, and rule-bending ideas are valid requests when you express them as deterministic in-world mechanics. Use features.events, features.commands, conditions, scoreboards, tags, particles, sounds, titles, messages, gamerules, time/weather actions, block changes, explosions, lightning, and relative teleports when those primitives faithfully match the user's requested behavior.
Do not silently reinterpret intent. If the declarative schema can faithfully express the request, use it. If it cannot, classify the request as needs_agent_code, generate the typed base project, implement the requested code, and submit it back through the agent-edited source testing loop. If a substitution is user-approved, set fidelity_policy="acknowledged_approximation" and still require the quality gate.
This deterministic builder does not generate arbitrary client mixins, shader loaders, native code, optimization core patches, or custom packet frameworks through the declarative schema. Those requests belong in the agent-edited source workflow, not in a closest-supported substitution presented as success.
Implementation truth
Comments, debug chat, README guidance, and metadata do not count as mechanics. Every accepted declarative action has a concrete platform implementation and a runtime assertion. A mechanic without that implementation rejects with MINECRAFT_CAPABILITY_NOT_IMPLEMENTED, including its exact field path and required_next_action="edit_source_and_start_build_job". There is no policy that converts incomplete behavior into a successful mod.
Features whose enabled_platforms exclude the selected target_platform are rejected with MINECRAFT_PLATFORM_FEATURE_MISMATCH; they are not dropped from the generated mod.
The request contract is pinned to schema_version="2026-05-05-v2" and compatibility_mode="strict". The former platform-passthrough mode was retired because it could report success after dropping features that excluded the selected platform.
Action-specific top-level fields are fail-closed. task_id belongs only to get_build_job; limit only to list_build_jobs; preview fields only to render_preview_image; and source_test_mode plus verification_contract require start_build_job with source_archive_file_id. Build/output/verification options are accepted only by create/build actions. Remove accidental cross-action fields instead of expecting them to be ignored.
Bedrock declarative features.dimensions is rejected even when experiments are enabled because the pinned Creator Tools cooperative-add-on gate rejects dimension definitions (CADDONIREQ191/CADDONREQ131). Use an edited standalone behavior-pack project and retain the full verification gate for dimension work; the builder does not return a known-invalid cooperative add-on.
Java client modules
Use features.client_modules for Java-only client utility behavior on target_platform="fabric" or target_platform="neoforge". Bedrock and skin packs reject client modules because they cannot install Java client code.
Supported module_kind values:
utility_client_preset — general anarchy/utility preset for requests phrased as "2b2t client", "anarchy client", or "hacked client"; implements an FPS HUD, bounded typed performance-option updates, maximum-valid brightness, and TNT-cart placement scanning.fps_hud — HUD-only FPS display.performance_profile — FPS-oriented HUD plus typed client option updates for render distance, simulation distance, entity distance scale, fast graphics, clouds off, and minimal particles.tnt_cart_placement_esp — client-side scanner for nearby rail blocks where TNT minecarts can be placed; renders a HUD count and nearest candidate coordinates.block_esp — bounded client scanner for listed block identifiers.entity_esp — bounded client scanner for listed entity identifiers.fullbright — sets the client gamma option to Minecraft's maximum valid brightness.
Common aliases are accepted for module_kind: hacked_client, anarchy_client, 2b2t_client, and utility_client normalize to utility_client_preset; shizik, fps_boost, and optimization_client normalize to performance_profile; tnt_esp, tnt_cart_esp, and minecart_tnt_esp normalize to tnt_cart_placement_esp. Renderer-hybrid aliases are recognized only so validation can return the edited-source next action; they do not generate metadata in place of renderer code.
Client-module options:
| field | Applies to | Notes |
|---|
scan_radius | ESP modules | Horizontal scan radius, 4 to 64; default 24. |
update_interval_ticks | all modules | Refresh interval, 1 to 200; default 10. |
max_rendered_markers | ESP modules | Maximum retained HUD markers, 1 to 128; default 32. |
render_style | ESP modules | Only hud is supported in this release; outline modes are rejected instead of ignored. |
color_hex | all modules | HUD/overlay accent color; accepts the same color forms as textures. |
identifiers | block_esp, entity_esp | Required list of block/entity ids to scan for. |
performance_mode | performance_profile, utility_client_preset | balanced, fps, or quality; default balanced. |
max_render_distance | performance modules | Applied render-distance option, 2 to 32; default 8. |
max_simulation_distance | performance modules | Applied simulation-distance option, 5 to 32; default 5. |
entity_distance_scale | performance modules | Applied entity render-distance scale, 0.1 to 1.0; default 0.5. |
Renderer rewrites require real loader-specific code and dependencies. Submit that edited Java source with a client verification contract; the declarative path rejects it instead of returning a misleading metadata-only artifact.
Examples:
{
"client_modules": [
{
"module_id": "anarchy_utility",
"display_name": "Anarchy Utility",
"module_kind": "2b2t_client",
"color_hex": "lime",
"scan_radius": 32,
"max_rendered_markers": 48
}
]
}
{
"client_modules": [
{
"module_id": "shizik_profile",
"display_name": "Shizik FPS Profile",
"module_kind": "shizik",
"performance_mode": "fps",
"max_render_distance": 6,
"max_simulation_distance": 5,
"entity_distance_scale": 0.35
}
]
}
{
"client_modules": [
{
"module_id": "tnt_cart_spots",
"display_name": "TNT Cart Placement ESP",
"module_kind": "tnt_cart_placement_esp",
"scan_radius": 48,
"update_interval_ticks": 5,
"max_rendered_markers": 64,
"color_hex": "#ff3355"
}
]
}
Versions, colors, and locales
Omit minecraft_version unless the user explicitly requires the pinned supported version. Each platform has one supported version in list_capabilities; unsupported older/newer versions are rejected instead of guessed. Whitespace is ignored, and v is accepted only when the remaining value exactly matches the pinned version.
Color fields (color_hex, brand_color_hex, hover_text_color, map_color, default_color) accept #RRGGBB, six-digit hex without #, three-digit short hex, or these color names: black, white, red, green, blue, yellow, orange, purple, pink, brown, gray, grey, cyan, magenta, lime, gold, silver. Values are normalized to lowercase #rrggbb.
Locales use ll_RR, such as en_US. Lowercase or hyphenated equivalents like en_us, en-us, and en-US are normalized.
Scoreboards, commands, functions, rarity
Use objective_id for scoreboards. Accepted aliases: scoreboard_id, objective. Use criterion; criteria is accepted.
Use description for commands. display_name is accepted as an alias when description is absent.
Functions may include optional description metadata. Use lines; commands is accepted as an alias when lines is absent.
Structured Bedrock functions are emitted under behavior_pack/functions/agentpmt/<mod_id>/. Advanced .mcfunction resources sent at older locations under behavior_pack/functions/ are moved into that same cooperative namespace, and exact references between uploaded functions are rewritten automatically so Creator Tools does not reject loose files.
features.functions[] requires at least one single-line Minecraft command in lines (or its commands alias); empty functions, multiline input, and shell commands reject before generation. Java features.tags[] likewise requires at least one valid resource id in values; short ids and short #tag references are namespaced to the current mod. Every features.localizations[] entry requires a non-empty translation mapping so the builder never emits an inert language file.
Use rarity.value for item rarity. rarity.rarity and rarity.name are accepted aliases.
Historical exact aliases also normalize when the intent is deterministic:
| Location | Canonical field/value | Accepted alias |
|---|
ItemSpec food field | saturation | top-level saturation_modifier |
ItemSpec bow-like item | item_kind="projectile" + shooter | tool_type="bow" |
BlockSpec.block_kind | basic | generic |
UISpec.ui_kind | hud_overlay | hud, overlay, screen_overlay |
UISpec.ui_kind | key_mapping | keybind, key_binding |
RecipeSpec.result_item | result_item | result, output, output_item, item_result, block_result, result_block, output_block, block_output |
ActionSpec aliases
Every action uses action_kind. particle_effect and spawn_particles are accepted as action_kind aliases for spawn_particle.
Broadcast aliases are accepted for message/title actions: broadcast_message, broadcast_chat, announce_message, and global_message normalize to send_message with audience="all_players"; broadcast_title, announce_title, and global_title normalize to show_title with audience="all_players".
When an action needs a target resource, prefer identifier. Accepted exact aliases:
| action_kind | Canonical field | Accepted aliases |
|---|
apply_effect, remove_effect | identifier | effect, effect_id |
spawn_entity | identifier | entity, entity_id |
play_sound | identifier | sound, sound_id |
spawn_particle | identifier | particle, particle_id |
add_tag, remove_tag | identifier | tag, tag_id |
give_item, drop_item | identifier | item, item_id |
give_item, drop_item | value | count, amount |
replace_held_item | identifier | item, item_id |
apply_damage, heal | value | amount |
set_block, place_block | identifier | block, block_id |
set_scoreboard | identifier | objective, objective_id, scoreboard_id |
set_scoreboard | |
Action options
Use these optional action fields only on the listed action_kind; unsupported combinations are rejected instead of ignored.
| action_kind | Supported options |
|---|
play_sound | volume (0 to 16), pitch (>0 to 4), position |
spawn_particle | count (1 to 256), spread_x, spread_y, spread_z (0 to 16), position |
spawn_entity, teleport, set_block, place_block, break_block_with_drops, create_explosion, summon_lightning | position |
apply_effect | amplifier (0 to 255) |
run_command | required command |
send_message, show_title | required message or scalar value text |
give_item, drop_item, apply_damage, heal, damage_item, repair_item | optional positive integer value |
modify_attribute | optional finite numeric value (defaults to 1) |
create_explosion | optional finite numeric value > 0 for power (defaults to 2) |
set_scoreboard | operation: set, add, or remove |
send_message, |
position is always a relative offset from the event target or event block, never an absolute world coordinate. Send it as {"x": x, "y": y, "z": z} or [x, y, z] with integer coordinates from -32 to 32.
set_entity_target, mount_entity, and tame_entity operate on the event target. Add a target_entity condition identifying a compatible mob. For backward compatibility, identifier, entity, or entity_id supplied on one of these actions inside an event is moved to that condition when it does not conflict with an explicit condition. Bedrock rejects set_entity_target because the stable Script API exposes the AI target as read-only. mount_entity and tame_entity reject targets that do not expose the required platform component/API.
For send_message and show_title, target is accepted only as an audience alias when it is one of source, self, player, @s, all, everyone, all_players, @a, or broadcast. Use canonical audience in new requests.
For give_item, drop_item, apply_damage, and heal, any provided value or accepted value alias must be an integer >= 1.
Generic action fields are fail-closed: identifier, value, command, message, and amplifier reject on action kinds that do not consume them. Remove an accidental field or use the canonical field for the intended action; the builder never silently drops it.
For set_scoreboard, operation defaults to set. Use add to increment and remove to decrement; both require a positive integer value or accepted value alias. Use the opposite operation instead of negative deltas.
Use wait inside event or Script API command action lists to delay all following actions. A wait must be followed by an executable action, consecutive waits must be combined, and chance_percent belongs on the following action or owning event rather than the wait. Bedrock .mcfunction commands cannot express wait or chance_percent; use simple_command or argument_command for those.
Custom command parameters
Use command_kind="argument_command" with parameters for runtime command input. Parameter names are referenced in action message or command strings as {param:name}.
Supported param_type values are string, integer, float, and player_name. A string parameter is greedy text and must be last. Optional parameters must come after all required parameters. Example:
{
"commands": [
{
"command_id": "announce",
"description": "Broadcast a title.",
"command_kind": "argument_command",
"permission_level": "op",
"parameters": [{"name": "message", "param_type": "string"}],
"actions": [
{
"action_kind": "show_title",
"audience": "all_players",
"title_display": "actionbar",
"message": "{param:message}",
"fade_in_ticks"
For spawn_particle, omitting count and spread fields preserves each platform generator's default particle behavior. Send explicit values when the count or spread matters.
ConditionSpec aliases
Every condition uses condition_kind. Prefer identifier; accepted exact aliases are item/item_id for held_item, entity/entity_id for target_entity, block for block_id, biome for biome_id, tag/tag_id for has_tag, and objective/objective_id/scoreboard_id for score_at_least. For score_at_least, score and amount are accepted aliases for value.
Condition value reference
entity_within_radius: value is an integer radius from 1 to 64; optional identifier filters the nearby entity type, for example {"condition_kind":"entity_within_radius","identifier":"minecraft:player","value":12}.time_of_day: value is one of day, night, dawn, sunrise, morning, noon, dusk, sunset, or midnight.weather_is: Java only; identifier is one of clear, rain, or thunder.y_below: value is an attainable Overworld threshold from -63 to 320; y_above accepts -64 to 319.light_level_below: Java only; value is an integer from 1 to 15 (0 is impossible because light levels are never negative).
Recipes
Each recipe is a minimal copy-pasteable spec. Drop the recipe inside the features block of a start_build_job request along with target_platform, mod_id, and mod_name.
Recipe 1 — A diamond sword
{
"items": [
{
"item_id": "flame_sword",
"display_name": "Flame Sword",
"item_kind": "weapon",
"tool_type": "sword",
"tool_tier": "diamond",
"damage": 7,
"texture": {"color_hex": "#ff5522"}
}
]
}
The validator confirms item_kind="weapon" with tool_type="sword". The renderer expands #ff5522 into a sword sprite. The item lands in the Equipment creative tab. Durability, repair material, enchantability, hand-equipped rendering, and the sword-tag default come from tool_tier="diamond".
Recipe 2 — An iron pickaxe
{
"items": [
{
"item_id": "iron_pickaxe",
"display_name": "Iron Pickaxe",
"item_kind": "tool",
"tool_type": "pickaxe",
"tool_tier": "iron",
"texture": {"color_hex": "#cccccc"}
}
]
}
tool_type="pickaxe" is required: tools without tool_type are rejected because the icon and dig speed would be non-deterministic. Default damage/durability come from tool_tier="iron". Substitute axe, shovel, or hoe for other vanilla tool kinds.
Recipe 3 — A custom-tier weapon
{
"items": [
{
"item_id": "void_blade",
"display_name": "Void Blade",
"item_kind": "weapon",
"tool_type": "custom",
"tool_tier": "custom",
"damage": 12,
"durability": 3000,
"tool": {
"tool_type": "custom",
"tier": "custom",
"attack_damage": 12,
"attack_speed": 1.4
},
"repairable": {"repair_items":
Use custom to opt out of tier defaults. Declare damage, durability, tool.attack_speed, repairable, and enchantable explicitly.
Recipe 3b — A pickaxe that mines every block
{
"items": [
{
"item_id": "godpick",
"display_name": "Godpick",
"item_kind": "tool",
"tool_type": "pickaxe",
"tool_tier": "netherite",
"tool": {
"tool_type": "pickaxe",
"tier": "netherite",
"break_all_blocks": true,
"instabreak": true
},
"texture": {"color_hex": "#ffd700"}
}
]
}
tool.break_all_blocks=true tells the generator: this tool ignores the vanilla pickaxe/axe/shovel/hoe destructible tags and mines every block. Bedrock emits a Molang-true tag predicate on minecraft:digger.destroy_speeds; Java upgrades the tier to netherite so the tool's INCORRECT_FOR block tag is effectively empty. Pair with instabreak: true for one-shot mining.
Recipe 4 — A full Java diamond armor set
{
"items": [
{"item_id": "set_helmet", "display_name": "Set Helmet", "item_kind": "armor", "tool_tier": "diamond", "armor": {"slot": "head", "protection": 3, "toughness": 2}, "texture": {"source_base64": "<base64 PNG of the helmet icon>"}},
{"item_id": "set_chest", "display_name": "Set Chest", "item_kind": "armor",
One ItemSpec per armor slot. armor.slot is required: head/chest/legs/feet. tool_tier must be iron, gold, diamond, or netherite and selects the vanilla armor material (sound, repair item, durability, and worn equipment asset). Supply a real PNG for the inventory icon; color_hex is rejected with MINECRAFT_ARMOR_TEXTURE_REQUIRES_IMAGE. Use edited source for wood/stone/custom armor materials or a custom worn atlas.
Recipe 5 — A food item
{
"items": [
{
"item_id": "spicy_jerky",
"display_name": "Spicy Jerky",
"item_kind": "food",
"food": {
"nutrition": 7,
"saturation_modifier": 0.6,
"effects": [{"effect": "minecraft:fire_resistance", "chance": 1.0, "duration_seconds": 30, "amplifier": 0}]
},
"texture": {"color_hex": "#7a3a1a"}
Pair item_kind="food" with the food preset. The food preset emits the eat animation, hunger restore, and any post-consumption effects.
Recipe 6 — A bow that fires arrows
{
"items": [
{
"item_id": "magic_arrow",
"display_name": "Magic Arrow",
"item_kind": "projectile",
"projectile": {"projectile_entity": "minecraft:arrow"},
"texture": {"color_hex": "#88aaff"}
},
{
"item_id": "magic_bow",
"display_name": "Magic Bow",
"item_kind": "projectile",
"shooter": {"charge_on_draw": true, "scale_power_by_draw_duration":
Two items: one for the projectile, one for the launcher. Both carry item_kind="projectile"; the launcher has the shooter preset.
Recipe 7 — An ore plus its worldgen and drop
{
"blocks": [
{
"block_id": "ruby_ore",
"display_name": "Ruby Ore",
"block_kind": "ore",
"hardness": 3.0,
"resistance": 8.0,
"drops": ["modid:ruby"],
"texture": {"color_hex": "#c91140"}
}
],
"items": [
{
"item_id": "ruby",
"display_name": "Ruby",
"item_kind": "generic",
"texture"
drops references an item id by namespace. The mined ruby is item_kind="generic" — it is a raw material that another recipe consumes; it does not need a tool/armor/food preset.
Recipe 8 — A hostile mob
{
"entities": [
{
"entity_id": "shadow_stalker",
"display_name": "Shadow Stalker",
"entity_kind": "hostile_mob",
"health": 30,
"attack_damage": 5,
"movement_speed": 0.32,
"spawn_biomes": ["minecraft:dark_forest", "minecraft:taiga"],
"spawn_weight": 12,
"texture": {"color_hex": "#1a1a2e"}
}
]
}
The entity texture seeds both the mob model and the spawn-egg color. Bedrock additionally accepts source_base64 or source_file_id to use a 64×64 PNG atlas.
Recipe 9 — A tameable pet
{
"entities": [
{
"entity_id": "stone_lynx",
"display_name": "Stone Lynx",
"entity_kind": "tameable_pet",
"health": 20,
"attack_damage": 4,
"tameable": {"tame_items": ["minecraft:cooked_beef"], "probability": 0.5},
"texture": {"color_hex": "#9d8866"}
}
]
}
entity_kind="tameable_pet" plus the tameable preset enables the right-click-with-item taming flow.
Recipe 10 — A weapon that sets enemies on fire
{
"items": [
{
"item_id": "flame_sword",
"display_name": "Flame Sword",
"item_kind": "weapon",
"tool_type": "sword",
"tool_tier": "diamond",
"damage": 8,
"texture": {"color_hex": "#ff5522"}
}
],
"events": [
{
"event_id": "flame_sword_ignite",
"event_kind": "on_item_hit_entity",
"conditions": [{"condition_kind":
One item, one event. The event filters by held_item (your sword's id) and emits set_on_fire against the hit entity. The event does not create a second item.
Recipe 11 — A skin pack with one skin
{
"action": "start_build_job",
"target_platform": "bedrock_skinpack",
"mod_id": "blue_pack",
"mod_name": "Blue Pack",
"skin_pack": {
"pack_id": "blue_pack",
"display_name": "Blue Pack",
"skins": [
{
"skin_id": "blue_hero",
"display_name": "Blue Hero",
"texture": {"source_base64": "<base64-encoded 64x64 PNG>"}
}
]
}
}
Skin pack textures must be exactly 64×64 PNG.
Recipe 12 — A cursed prank item
{
"items": [
{
"item_id": "cursed_stick",
"display_name": "Cursed Stick",
"item_kind": "generic",
"texture": {"color_hex": "#7a35ff"}
}
],
"events": [
{
"event_id": "cursed_stick_use",
"event_kind": "on_item_use",
"conditions": [{"condition_kind": "held_item", "identifier": "cursed_stick"}],
"actions": [
{
Use relative position for harmless movement effects. The item stays one ItemSpec; the event supplies the prank behavior.
Recipe 13 — A fake-glitch block event
{
"blocks": [
{
"block_id": "glitch_panel",
"display_name": "Glitch Panel",
"block_kind": "interactive",
"texture": {"color_hex": "#00ffff"}
}
],
"events": [
{
"event_id": "glitch_panel_interact",
"event_kind": "on_block_interact",
"conditions": [{"condition_kind": "block_id", "identifier": "glitch_panel"}],
"actions": [
{
Fake-glitch behavior should be built from world-side events, particles, titles/messages, time, weather, sounds, and block changes.
Recipe 14 — An admin debug command
{
"commands": [
{
"command_id": "debug_storm",
"description": "Toggle a dramatic debug storm.",
"command_kind": "admin_command",
"permission_level": "admin",
"actions": [
{"action_kind": "set_gamerule", "identifier": "doDaylightCycle", "value": false},
{"action_kind": "set_time", "value": "night"},
{"action_kind": "set_weather", "identifier": "thunder"
Use admin_command plus permission_level="admin" for operator tools. Prefer set_gamerule, set_time, and set_weather over raw command strings when those first-class actions cover the behavior.
Recipe 15 — A chaos trap
{
"blocks": [
{
"block_id": "chaos_plate",
"display_name": "Chaos Plate",
"block_kind": "interactive",
"texture": {"color_hex": "#ff00ff"}
}
],
"scoreboards": [{"objective_id": "chaos_triggers", "display_name": "Chaos Triggers", "criterion": "dummy"}],
"events": [
{
"event_id": "chaos_plate_trigger",
"event_kind": "on_block_interact",
Keep high-impact effects bounded with explicit values and relative positions so the generated behavior remains predictable.
Recipe 16 — Timed/random progression with milestones
{
"scoreboards": [{"objective_id": "weeks", "display_name": "Weeks", "criterion": "dummy"}],
"events": [
{
"event_id": "weekly_progress",
"event_kind": "on_player_tick",
"interval_ticks": 1200,
"cooldown_ticks": 1200,
"chance_percent": 35,
"actions": [
{"action_kind": "set_scoreboard", "identifier": "weeks", "operation": "add"
Use declared scoreboards for staged Java progress; the objective is created on server startup and behavior verification asserts that it exists. set_scoreboard can also create an undeclared writable objective at action time on both Java and Bedrock. A declared objective mutated by set_scoreboard must use criterion="dummy" or "trigger"; read-only vanilla criteria reject before generation. Use interval_ticks for cadence, cooldown_ticks to prevent repeated firing, and chance_percent for random outcomes.
Recipe 17 — A stalker entity with proximity behavior
{
"entities": [
{
"entity_id": "stalker",
"display_name": "Stalker",
"entity_kind": "hostile_mob",
"health": 60,
"movement_speed": 0.35,
"scale": 2.5,
"ai_goals": [
{"goal_kind": "nearest_attackable_target", "priority": 1, "search_radius": 30},
{"goal_kind": "melee_attack", "priority": 2, "speed_multiplier": 1.25
attack_reach_multiplier is Bedrock-only. Fabric and NeoForge reject it with an actionable platform error; use search_radius and radius actions for cross-platform sensing.
Recipe 18 — A two-input machine
{
"items": [
{"item_id": "amalgam", "display_name": "Amalgam", "texture": {"color_hex": "#b8f2ff"}}
],
"machines": [
{
"machine_id": "amalgamator",
"display_name": "Amalgamator",
"texture": {"color_hex": "#606060"},
"processing_time_ticks": 100,
"machine_recipes": [
{
"recipe_id": "make_amalgam",
"inputs":
Right-click the generated machine with each input item. Inputs are tracked per block position; when the full set is present, the output drops after processing_time_ticks.
Verify Every Generation
Use render_preview_image to inspect important textures, then submit the same spec through start_build_job. A visual preview checks the selected texture only; the async quality gate proves packaging, runtime loading, declared behavior, and client/render checks before installation.
{
"action": "render_preview_image",
"target_platform": "bedrock",
"mod_id": "flame_tools",
"mod_name": "Flame Tools",
"preview_target_kind": "item",
"preview_target_id": "flame_sword",
"features": { }
}
If the preview looks like a flat colored square or the wrong tool sprite, you forgot item_kind, tool_type, or both.
Common Mistakes
Each entry pairs the failure mode with the structured error code the validator now returns. If you see the code, fix the spec.
| Mistake | What happens now |
|---|
item_kind="generic" on a weapon with damage set | Rejected: MINECRAFT_ITEM_KIND_REQUIRES_PRESET |
item_kind="tool" without tool_type | Rejected: MINECRAFT_ITEM_KIND_REQUIRES_PRESET |
item_kind="weapon" with tool_type="pickaxe" | Rejected: weapon requires sword or custom |
item_kind="armor" without armor or wearable | Rejected: MINECRAFT_ITEM_KIND_REQUIRES_PRESET |
item_kind="food" without food preset or nutrition | Rejected: MINECRAFT_ITEM_KIND_REQUIRES_PRESET |
item_kind="block_placer" without block_placer | Rejected: MINECRAFT_ITEM_KIND_REQUIRES_PRESET |
food preset on an item_kind="tool" | Rejected: preset/kind mismatch |
texture omitted on any visible asset | Rejected: MINECRAFT_VISIBLE_ASSET_TEXTURE_REQUIRED |
texture: {} with no color_hex, source_base64, or source_file_id | Rejected: MINECRAFT_TEXTURE_SOURCE_REQUIRED |
item_kind="armor" with texture: {color_hex: ...} only | Rejected: MINECRAFT_ARMOR_TEXTURE_REQUIRES_IMAGE |
| Pickaxe specced to "break any block" but only breaks a few | Bedrock: add tool.break_all_blocks: true. Fabric/NeoForge: use the agent-edited source workflow; declarative requests reject because changing to a vanilla netherite tier does not make every block mineable. |
| Feature appears in the wrong platform build | Set enabled_platforms: ["bedrock"] (or fabric/neoforge) to scope it; an empty list applies to all platforms |
Supported Platforms
bedrock: Minecraft Bedrock add-on generation for pinned Mojang Bedrock Samples baseline v1.26.20.4. Produces source zips, .mcaddon installables, and Mojang Minecraft Creator Tools validation reports.fabric: Minecraft Java Fabric generation for pinned version 1.21.11. Produces source zips and optional built jars using the pinned offline Gradle/Fabric toolchain.neoforge: Minecraft Java NeoForge generation for pinned Minecraft 1.21.11 and NeoForge 21.11.42. Produces source zips and optional built jars using the pinned offline Gradle/NeoForge toolchain.bedrock_skinpack: Bedrock skin pack generation. Produces installable .mcpack files from skin texture images and localization/manifest metadata.forge is retired. Requests using target_platform="forge" are rejected with a structured migration error directing callers to neoforge.
Actions
list_capabilities
Return supported platforms, pinned dependencies, feature kinds, output artifacts, and platform notes.
Optional:
target_platform: bedrock, bedrock_skinpack, fabric, or neoforge.
validate_mod_project
Validate either a structured mod specification or a source archive previously produced by this tool.
Modes:
- Spec validation: provide
target_platform, mod_id, mod_name, and features (or skin_pack for skinpacks).
- Archive validation: provide
source_archive_file_id.
create_mod_project
Generate synchronous source/debug artifacts without runtime verification. This action accepts only verification_level="off", always returns ready_for_install=false, and must not be used for final delivery.
Required:
target_platform: bedrock, bedrock_skinpack, fabric, or neoforge.mod_id: lowercase identifier matching ^[a-z][a-z0-9_]{1,63}$.mod_name: human-readable mod name.features: structured feature object for normal Bedrock/Fabric/NeoForge mods.skin_pack: structured skin pack object for target_platform=bedrock_skinpack.
Optional:
minecraft_version: omit to use the pinned platform version.description: short mod description.output_mode: installable, source, or both. Default is both.build_jar: for Fabric/NeoForge only. Default is true.validate_output: run external platform validation where available. Default is true.include_file_preview: include capped text previews in generated_files.allow_experimental_bedrock_features: allow script-based Bedrock features that may require experiments.authoring_mode: auto or declarative for typed features. agent_code is accepted only by start_build_job with source_archive_file_id.
start_build_job, get_build_job, and list_build_jobs
Use start_build_job for final artifacts. It queues compile/package/runtime work and immediately returns a task_id. Poll get_build_job with that id until status is completed or failed; inspect result.ready_for_install, result.quality_gate, and result.runtime_verification. list_build_jobs returns recent jobs for the current budget.
For generated specs, provide the same required fields as create_mod_project. For agent-edited source, provide authoring_mode="agent_code", source_archive_file_id, target_platform, mod_id, and source_test_mode. verification_level="behavior" also requires a machine-checkable verification_contract bound to the same platform and mod id with at least one behavior, render, or client check; boot_smoke may omit it because that level proves loading only, never behavior.
render_preview_image
Render an enlarged PNG preview for an item, block, or entity texture and upload it to File Manager. Use it before the final verified build to confirm important icons look right.
Modes:
- Spec preview: provide
target_platform, mod_id, mod_name, and features; the tool generates files in memory and previews the selected texture.
- Source archive preview: provide
source_archive_file_id from a generated source zip; optionally include mod_id.
- Direct image preview: provide
preview_source_file_id for a File Manager image uploaded by the user.
Optional:
preview_target_kind: item, block, or entity.preview_target_id: feature id or namespaced id, such as flame_sword or flame_tools:flame_sword.preview_size: square preview PNG size from 32 to 1024 pixels. Default is 256.preview_background: checkerboard, transparent, white, or black. Default is checkerboard.
Field Reference
All feature arrays are optional. Use empty arrays or omit unused sections.
Texture binding (BindingTextureSpec)
Every visible asset's texture field is a BindingTextureSpec. Supply exactly one of:
source_base64: base64-encoded image bytes (PNG or JPEG, max 1024 px, max 5 MB).source_file_id: File Manager file_id for a user-uploaded image.color_hex: explicit #RRGGBB color rendered as a kind-shaped procedural sprite.
source_base64 and source_file_id are mutually exclusive. A spec with none of the three is rejected with MINECRAFT_TEXTURE_SOURCE_REQUIRED.
Items
item_id: required lowercase identifier.display_name: required.item_kind: required behavior class (see Recipes). One of generic, tool, weapon, armor, food, fuel, projectile, block_placer, or entity_placer. Drives icon sprite, creative-tab placement, stack size, and required presets.tool_type: pickaxe, axe, shovel, hoe, sword, or custom. Required for item_kind="tool"; weapon requires sword or custom.tool_tier: wood, stone, iron, gold, diamond, netherite, or custom for tools/weapons. Armor permits only iron, gold, diamond, or netherite and defaults to iron; unsupported armor tiers reject instead of silently becoming iron.texture: required BindingTextureSpec.damage, durability, max_stack_size, nutrition, saturation: per-field semantics; defaults derive from item_kind/tool_tier.tool, armor, food, block_placer, entity_placer, projectile, shooter, throwable: presets used by the matching item_kind. The validator rejects preset/kind mismatches with MINECRAFT_ITEM_KIND_REQUIRES_PRESET. Bedrock honors tool.mining_speed, instabreak, and break_all_blocks; Java custom mining rules, nested digger/repair/enchant presets, and Bedrock custom attack speed require edited source and reject declarative generation.
Blocks
block_id: required lowercase identifier.display_name: required.block_kind: basic, ore, light, interactive, crop, plant, redstone, machine, or container. Consult list_capabilities.implementation_status for the target: Java rejects redstone because the declarative Java block does not implement signal behavior, while Bedrock implements conductor behavior but rejects signal emission.texture: required BindingTextureSpec.- Implemented fields depend on the selected block kind and platform. Shared declarative fields include
hardness, resistance, light_level, drops, and texture; Bedrock supports the typed component fields accepted by its pinned registry, including collision_box, selection_box, placement_filter, flammable, friction, map_color, crop, redstone, tick, crafting_table, and component_overrides. Explicit material_instances and custom_components require edited source because the declarative renderer derives the material instance from texture and does not register custom component code.
Entities
entity_id: required lowercase identifier.display_name: required.entity_kind: passive_mob, hostile_mob, neutral_mob, projectile, npc_like, tameable_pet, rideable_mount, or boss_like. Java declarative generation accepts the first four; Java requests for the latter four reject and route to edited source. Bedrock also rejects npc_like: the old path emitted a fixed apple trade rather than the requested trade contract. Use edited source for NPC trades; other Bedrock kinds are accepted only when every populated behavior field has a real pack implementation. ai_goals[].goal_kind=trade_with_player and every interactions[] preset likewise reject until an explicit trade/interaction implementation is supplied through edited source.texture: required BindingTextureSpec. Seeds both the mob model and the spawn-egg color.health, attack_damage, movement_speed, spawn_biomes, spawn_weight, families, ai_goals, interactions, rideable, tameable, breedable, equipment, render, loot_table, component_overrides.
Particles
particle_id: required lowercase identifier.texture: required BindingTextureSpec.
Bedrock custom particles generate a resource-pack particle definition and a textures/agentpmt/<mod_id>/particle/{particle_id}.png sprite. Spawn a custom particle with {"action_kind": "spawn_particle", "identifier": "<particle_id>"} or the legacy <mod_id>:<particle_id> form; the emitted Bedrock pack uses creator_project:<particle_id>. Use vanilla identifiers such as minecraft:basic_flame_particle when a custom texture is not needed.
Bedrock animations and controllers
Animation resources are Bedrock-only and must contain executable content. AnimationSpec.animation_json requires a non-empty animations object with at least one bones, sound, particle, or timeline behavior. AnimationControllerSpec.controller_json requires a state with an animation, transition, or event. RenderControllerSpec.controller_json requires a controller with non-empty geometry, materials, and textures. Empty resources are rejected instead of being emitted as inert placeholders.
Effects
Declarative custom effect definitions are rejected on every platform because the former generated registrations did not implement their advertised mechanics. To apply or remove a real vanilla effect, use the event/command actions apply_effect and remove_effect with a minecraft:<effect> identifier. Implement a genuinely custom status effect in agent-edited source and submit it through start_build_job with behavior checks.
Commands
command_id: required lowercase identifier.description: required.command_kind: simple_command, argument_command, function_command, or admin_command.permission_level: any, op, or admin.actions: required non-empty array of action specs.
Worldgen
feature_id: required lowercase identifier.feature_kind: ore_feature, single_block_feature, vegetation_patch, structure_placement, custom_biome, or custom_dimension.block_id, target_block (default minecraft:stone), vein_size, count_per_chunk, min_y, max_y, biomes.
min_y must be less than or equal to max_y.
Recipes
recipe_id: required lowercase identifier.result_item: required item or block item identifier. Accepted aliases: result, output, output_item, item_result, block_result, result_block, output_block, block_output.recipe_kind: shaped, shapeless, furnace, brewing_mix, brewing_container, smithing_transform, or smithing_trim.pattern, key, ingredients, count, cooking_time_seconds, input_item, template_item, base_item, addition_item.
Java targets reject brewing_mix and brewing_container with MINECRAFT_COMPONENT_UNSUPPORTED_ON_PLATFORM; those recipe kinds are Bedrock-only.
Loot Tables
Bedrock features.loot_tables entries must set target to one generated entity id. The generator writes the table under behavior_pack/loot_tables/agentpmt/<mod_id>/ and binds that exact path to the entity's minecraft:loot component. EntitySpec.loot_table may name the declared loot_id; legacy loot_tables/.../<loot_id>.json values resolve by the final filename. Conflicting, duplicate, undeclared, or missing entity targets are rejected. Java declarative loot tables remain rejected because generated Java entities do not yet bind per-entity death loot; use edited source there. For block drops on every platform, use BlockSpec.drops.
Events
event_id: required lowercase identifier.event_kind: one of the supported kinds per platform (see Recipe 10).conditions: optional array of condition specs (held_item, target_entity, source_entity, block_id, biome_id, has_tag, score_at_least, entity_within_radius, time_of_day, weather_is, y_below, y_above; Java also supports light_level_below).held_item, target_entity, and source_entity require the matching resource identifier; source_entity accepts entity and entity_id as aliases. score_at_least requires an objective identifier and finite integer value, and checks the source player's score on every platform. has_tag.subject is source or target and defaults to target.- Condition values:
entity_within_radius.value is 1 to 64 and its optional identifier filters entity type; time_of_day.value is day, night, dawn, sunrise, morning, noon, dusk, sunset, or midnight; weather_is.identifier is Java-only clear, rain, or thunder; y_below.value is -63 to 320; y_above.value is -64 to 319; light_level_below.value is Java-only 1 to 15. Java source_entity accepts only because Java event dispatch exposes the triggering player as source. Condition fields reject on kinds that do not consume them.
on_enter_biome fires on the player's first observed biome and only when that biome id changes afterward; add a biome_id condition to select a specific destination biome. on_item_inventory_tick scans every occupied inventory slot and requires a held_item condition identifying the stack. on_inventory_change evaluates held_item against the changed stack. Slot-mutating consume_item, damage_item, and repair_item actions are rejected on both inventory event kinds because declarative actions mutate the main hand rather than an arbitrary inventory slot; implement explicit slot mutation through the agent-edited source workflow.
Target-mutating actions reject on on_entity_death, where the target is no longer live. On on_projectile_hit, target-mutating actions require a non-player target_entity condition so the entity-hit branch cannot accidentally apply the action to the projectile or shooter after a block impact. Event cooldowns are per source player for player/source events and per target entity for entity events.
run_command accepts a single Minecraft command line only. Shell-like syntax such as &&, ||, backticks, $(), redirection, and shell/program names such as bash, npm, gradle, java, or python are rejected.
Machines and Storage
MachineSpec requires texture (visible block face). The declarative runtime implements machine_kind="processor"; generator, storage, tank, and transfer_endpoint reject and route to edited source.processing_time_ticks: ticks between completing the input set and producing output. processing_time is accepted as an alias; processing_time_seconds is accepted when it converts to whole ticks.machine_recipes: inline machine recipes. Each recipe has recipe_id, inputs (1-9 item ids), output, and output_count.- Machine recipes are interact-driven: right-click the machine with each required input item. The generated machine records pending inputs per block position, consumes matching inputs, waits
processing_time_ticks, then drops the output item.
- Bedrock machine textures are merged into the normal
resource_pack/textures/terrain_texture.json atlas. Do not remove the structured machine helper to avoid loose texture files.
StorageSpec declaratively supports a Bedrock backpack with item slots. Java storage and Bedrock item_container, fluid_tank, and energy_storage requests reject and route to edited source instead of emitting inert capacity metadata.
Skins
SkinPackSpec.skins[] requires skin_id, display_name, and a 64×64 texture.
Output Behavior
- Bedrock
installable or both returns a .mcaddon and validation report. With validate_output=true, Mojang Minecraft Creator Tools validates the generated add-on.
- Bedrock skin packs return a
.mcpack; with validate_output=true, the service first validates the archive structure, references, PNG integrity, and exact 64x64 skin dimensions. Minecraft Creator Tools then runs its supported skin checks, and only the add-on suite's expected missing-BP/RP-manifest findings (CADDONREQ163 and CADDONREQ164) are ignored; every other Creator Tools error remains fatal.
- Fabric/NeoForge
installable or both returns a jar when build_jar=true; otherwise use source for source zips only.
start_build_job returns a task_id immediately. get_build_job returns queued/processing/failed/completed state; completed jobs expose artifacts, validation/build output, runtime checks, the quality gate, and install readiness.render_preview_image returns a preview artifact with a PNG file_id, signed URL, and image metadata. It previews texture/icon assets only; start_build_job performs gameplay verification.- Supported Fabric/NeoForge entity kinds emit registered
EntityType definitions, concrete generated entity subclasses, default attributes, spawn eggs, client renderer registrations, and loader-specific attribute hooks. Unsupported kinds or populated fields reject before generation.
- Supported Fabric/NeoForge
ore_feature and single_block_feature worldgen specs emit configured/placed features plus loader-specific biome hooks. Other worldgen kinds reject before generation.
on_item_hit_entity event specs emit loader-specific attack-event handlers and execute supported Java action templates such as fire, damage, healing, effects, particles, sounds, tags, item grants/drops, entity spawns, and Minecraft commands.- Java builds use a per-job Gradle home and immutable pinned offline dependency/loader caches. No dependency versions are chosen at runtime and uploaded code cannot mutate the shared cache.
- Artifacts expire according to the File Manager storage policy. Store returned
file_id values if another tool needs to retrieve or validate them later.
Maintenance
- Regenerate
minecraft_mods_service/bedrock_component_registry.json after updating the pinned Mojang samples archive with tools/regenerate_bedrock_component_registry.py.
- Regenerate the public JSON schema after editing
minecraft_mods_service/models.py with tools/regenerate_public_tool_schema.py. The regenerated artifact is written to two locations and checked for byte-equality by test_schema_single_source.py.
When To Use
- Use this skill for
Minecraft Custom Mod Builder on AgentPMT.
- Use it when an agent needs this specific tool's behavior, schema, inputs, outputs, and invocation shape.
- Search and activation keywords: minecraft custom mod builder, create runtime verified bedrock add ons and skin packs; build and verify fabric and neoforge jars; generate typed items, blocks, entities, commands, create mod project, target platform, mod id.
- Supported action names:
create_mod_project, get_build_job, get_instructions, list_build_jobs, list_capabilities, render_preview_image, start_build_job, validate_mod_project.
Use Cases
- Create runtime-verified Bedrock add-ons and skin packs; build and verify Fabric and NeoForge jars; generate typed items
- blocks
- entities
- commands
- events
- recipes
- particles
- sounds
- worldgen
- scoreboards
- processor machines
- storage
- and supported client modules; review File Manager visual_proof galleries before delivery; preview visible assets; download editable source; upload agent-edited source for isolated compile/package/server/client testing; detect broken
- unverified
- or visually inadequate artifacts before installation
Related Product Skills
Categories And Industries
No categories or industry tags are published for this tool.
Actions And Schema
Complete generated action schema: ./schema.md.
Supported action count: 8.
x402 action routes are enabled and listed in ./schema.md.
create_mod_project (action slug: create-mod-project): Synchronous unverified source/debug generation. Accepts only verification_level='off'; use start_build_job for install-ready final artifacts. Price: 15 credits. Parameters: advanced_resources, allow_experimental_bedrock_features, assets, authoring_mode, build_jar, compatibility_mode, description, features, plus 12 more.get_build_job (action slug: get-build-job): Fetch one tenant-scoped build job by task_id. Poll until status is completed or failed, then inspect result.runtime_verification, result.visual_proof (File Manager frame file_ids/signed_urls; also in artifacts with label=visual_proof_frame), result.quality_gate, result.ready_for_install, and artifacts. Open visual proof frames before customer delivery. Price: 2 credits. Parameters: task_id.get_instructions (action slug: get-instructions): Get tool instructions and available actions. Price: 15 credits. Parameters: none.list_build_jobs (action slug: list-build-jobs): List recent Minecraft build jobs for the active budget to recover task_id values or inspect recent queued/completed builds. Price: 2 credits. Parameters: limit.list_capabilities (action slug: list-capabilities): Return supported platforms, pinned versions, feature/action matrices, implementation_status, quality-gate fields, and unsupported families. Price: 2 credits. Parameters: target_platform.render_preview_image (action slug: render-preview-image): Render and upload an enlarged PNG preview for an item, block, or entity from a structured spec, source archive, or direct image file. Price: 5 credits. Parameters: features, mod_id, mod_name, preview_background, preview_size, preview_source_file_id, preview_target_id, preview_target_kind, plus 3 more.start_build_job (action slug: ): Queue generated or agent-edited source for compile, package, real Minecraft runtime checks, File Manager visual_proof gallery, and the install-readiness gate. Poll get_build_job and ship only when result.ready_for_install=true and quality_gate.status='passed', after reviewing visual_proof frames. Price: credits. Parameters: , , , , , , , , plus 15 more.
Live Schema And Examples
Use the compact schema above for ordinary calls. Before a new production integration, or whenever parameters, enum values, nested objects, outputs, or examples are unclear, fetch live details first.
- Exact schema: call
agentpmt-tool-search-and-execution with action: "get_schema", and tool_id: "minecraft-custom-mod-builder".
- Detailed examples: call
agentpmt-tool-search-and-execution with action: "get_instructions" and tool_id: "minecraft-custom-mod-builder", or call this product with action: "get_instructions" when the product tool is already selected.
- Treat returned live schema and instructions as more specific than this generated summary.
MCP schema lookup through the main AgentPMT MCP server:
{
"method": "tools/call",
"params": {
"name": "AgentPMT-Tool-Search-and-Execution",
"arguments": {
"action": "get_schema",
"tool_id": "minecraft-custom-mod-builder"
}
}
}
For live examples, keep the same MCP tool and use these arguments:
{
"action": "get_instructions",
"tool_id": "minecraft-custom-mod-builder"
}
Authenticated AgentPMT REST schema lookup body:
{
"name": "agentpmt-tool-search-and-execution",
"parameters": {
"action": "get_schema",
"tool_id": "minecraft-custom-mod-builder"
}
}
Authenticated AgentPMT REST live examples body:
{
"name": "agentpmt-tool-search-and-execution",
"parameters": {
"action": "get_instructions",
"tool_id": "minecraft-custom-mod-builder"
}
}
Call This Tool
Product slug: minecraft-custom-mod-builder
Marketplace page: https://www.agentpmt.com/marketplace/minecraft-custom-mod-builder
- AgentPMT account route: first use
../agentpmt-account-mcp-rest-api-setup to connect the main MCP server or REST API for an Agent Group where this tool is enabled.
- No-account AgentAddress/x402 route: first use
../agentpmt-no-account-agentaddress-x402 for the canonical payment and wallet setup instructions.
- AgentPMT overview: use
../what-is-agentpmt for marketplace, Agent Group, workflow, MCP, REST, and payment concepts.
If those setup skills are not installed beside this product skill, use the downloads below.
Core AgentPMT setup skills:
- What AgentPMT is: ../what-is-agentpmt
- AgentPMT account MCP/REST setup: ../agentpmt-account-mcp-rest-api-setup
