Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add lightblink/cocos-asset-forge-mcp --skill cocos-asset-pipeline-directorgit clone --depth 1 https://github.com/lightblink/cocos-asset-forge-mcpWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/lightblink/cocos-asset-forge-mcp/cocos-asset-pipeline-director)<a href="https://agentmods.dev/skills/lightblink/cocos-asset-forge-mcp/cocos-asset-pipeline-director"><img src="https://agentmods.dev/badge/skills/lightblink/cocos-asset-forge-mcp/cocos-asset-pipeline-director/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/lightblink/cocos-asset-forge-mcp/cocos-asset-pipeline-director"><img src="https://agentmods.dev/badge/skills/lightblink/cocos-asset-forge-mcp/cocos-asset-pipeline-director.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00091 | $0.03216 |
| Opus 5 | $0.00046 | $0.01608 |
| Sonnet 5 | $0.00018 | $0.00643 |
| Haiku 4.5 | $0.00009 | $0.00322 |
Grade A, and why
cocos-asset-pipeline-director scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 12d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 208 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Cocos Asset Pipeline Director
Overview
Use this skill to direct Cocos Creator asset work through a dedicated asset-generation MCP instead of treating generated art as loose files. The skill chooses the right asset pipeline, constrains prompts for game-runtime use, preserves generated metadata, and verifies that outputs are ready to import or wire into a Cocos project.
The skill is intentionally a pipeline layer, not a competing image model wrapper. If the configured MCP can produce or adapt the requested asset, use it first.
Use cocos-asset-review-director after generation or adaptation when assets will be imported into gameplay, especially for player/enemy sprites, sprite sheets, VFX, UI packs, audio cues, directional or semantic art, or anything whose readability affects the first loop.
Compatibility
Expected MCP capabilities:
asset_forge_get_configasset_forge_plan_packasset_forge_generate_spriteasset_forge_generate_sprite_sheetasset_forge_generate_sprite_grid_sheetasset_forge_generate_tilesetasset_forge_generate_ui_packasset_forge_adapt_imageasset_forge_generate_sfxasset_forge_generate_music_loopasset_forge_adapt_audio
If one of these tools is unavailable, use the closest available MCP tool and clearly report the gap. Only use generic image generation, file downloads, or hand-authored placeholder assets when the MCP cannot satisfy the request or the user explicitly asks for another path.
Core Workflow
- Inspect configuration.
- Call
asset_forge_get_configbefore a non-trivial asset task. - Note the default output directory, Cocos version target, asset root, overwrite policy, and exposed providers.
- Call
- Clarify the runtime job.
- Identify whether the asset is a placeholder, production candidate, UI element, gameplay sprite, animation, tileset, effect, or imported external source.
- Identify target platform constraints: mobile, web, mini-game, playable ad, desktop, or editor-only prototype.
- For a from-zero playable slice, define the minimum asset pack before implementation starts: gameplay placeholders, UI/HUD art, feedback/VFX cues, SFX, and optional short music loop. If the slice intentionally uses engine geometry only, record that as a deliberate prototype constraint instead of silently skipping the asset pipeline.
- Apply a batch-first cost policy for generated images: related sprites, states, VFX frames, icons, pickups, enemy variants, projectiles, and props should be generated as one or more grid/contact sheets and sliced into frames. Use one-by-one sprite generation only for a style anchor, a uniquely important hero asset, or a failed cell that cannot be adapted.
- Establish a compact art bible.
- Capture camera/view, palette, outline style, rendering style, scale, silhouette rules, UI material, lighting, and negative prompts.
- Reuse existing canonical assets when present instead of regenerating known characters or UI styles.
- Select the MCP pipeline by runtime job.
- Generate into the MCP output directory unless the user provides a project-specific output path.
- Review outputs before import.
- For gameplay-critical sprites and VFX, apply an explicit asset contract: role, gameplay meaning, runtime size, motion requirement, style constraints, alpha policy, collision expectation, and import target.
- Use
cocos-asset-review-directorfor design fit, gameplay semantics, style consistency, animation readiness, alpha/silhouette issues, mobile thumbnail readability, collision fit, audio fit, and performance risk. - Do not import failed production-candidate assets silently. Adapt or regenerate the smallest failed unit.
- Import or stage assets under Cocos-friendly project folders only when requested or clearly required by the task.
- Report exact paths, tool choices, generation settings, review status, and any rejected or deferred assets.
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 12d ago First seen · 208 lines · 91 tokens per session scan A 1230081361fa
cocos-asset-pipeline-director is a skill published in the GitHub repository lightblink/cocos-asset-forge-mcp (1 stars, last pushed 2mo ago), licensed MIT. It adds 91 tokens to every session and 3,216 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
cocos-scene-prefab-assembly
Assemble Cocos Creator 3.8 scenes, prefabs, nodes, components, serialized properties, UI roots, managers, pools, and asset references into playable local game scenes. Use whenever the user asks Codex to wire gameplay into a Cocos scene, create or update prefabs, attach scripts, assign generated assets, define node…
cocos-wechat-local-build
Build, inspect, and debug local Cocos Creator 3.8 WeChat Mini Game packages without publishing. Use whenever the user asks Codex to prepare a Cocos project for local wechatgame build, create or review build configs, run Cocos Creator command-line builds, inspect build/wechatgame output, check package budgets, or…
cocos-creator-gameplay-architecture
Implement and review Cocos Creator 3.8 gameplay architecture, TypeScript components, UI systems, scene wiring, prefabs, resource loading, input, audio, tweening, collision, object pooling, state machines, and maintainable runtime code. Use whenever the user asks Codex to build gameplay, create a Cocos prototype…
cocos-game-design-director
Direct production-quality game design for Cocos Creator and WeChat Mini Game projects. Use when Codex needs to turn a game idea into a playable concept, define a core loop, scope an MVP or vertical slice, critique mechanics, write a design brief, choose player goals and rewards, or align gameplay implementation with…
cocos-game-reference-research-director
Research and deconstruct reference games, genre expectations, player promises, mechanics, content patterns, monetization-adjacent systems, art/audio cues, UX conventions, and IP-safe design pillars before planning or implementing Cocos Creator and WeChat Mini Game projects. Use when the user describes a game by…
cocos-interaction-ux-director
Design and critique touch-first interaction, game UX, UI states, onboarding, feedback, and control feel for Cocos Creator and WeChat Mini Game projects. Use when Codex needs to improve controls, menus, HUD, restart flow, tutorials, input mapping, moment-to-moment feedback, or player comprehension before or during…