anime-forge

AI 生成动画短片生产线：图（关键帧锁确定性）→ 视频（i2v 给活气）→ 后期（剪辑做节奏）。 Use when: 做动画短剧/角色 IP 短片、猫咖日记系列新集、图生视频管线、剧情类 AI 视频。 Not for: 录屏/教程/showcase 视频（用 video-forge）、单张图生成（用 image-generation）、PPT（用 ppt-forge）。 Output: charter+分镜+prompt book+素材账本+EDL 驱动的成片，全程可复现可重渲。

Quality

88%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Quality

Content

77%Weight 40%Scale 1-3

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A dense, highly actionable skill body with a well-sequenced workflow and real validation checkpoints. The main gaps are inline time-sensitive provenance dates that hurt conciseness, and a set of clearly-signaled references that are not backed by an actual shipped bundle.

Suggestions

Move provenance dates and version numbers (e.g. '2026-06-10~12', '2026-06-12 实测', 'v0.1') into a dedicated changelog/provenance or 'old patterns' section so the main flow stays time-insensitive and leaner.

Ship the referenced files (prompt-book-v0.1.md, review-protocol-v0.1.md, build-animatic.mjs, shot-plan-v0.1.md, assets/README.md) in references/ or scripts/ so the one-level-deep pointers resolve to real bundled material.

For references that point to external repo paths (docs/videos/cucu-pr-flow/episode-brief.md, video-forge/SKILL.md), note explicitly that they are cross-skill/repo artifacts rather than bundled references, so navigation expectations are clear.

Dimension	Reasoning	Score
Conciseness	The body is dense and un-padded ('不教什么是 i2v'), but inline time-sensitive dates ('2026-06-10~12', '2026-06-12 实测') and version numbers ('v0.1') sit in the main flow rather than a deprecated/old-patterns section, which the guidelines say to penalize — so not the lean level-3 anchor.	2 / 3
Actionability	Concrete executable guidance throughout: tool routing ('GPT 系生图', 'Gemini 系'), a specific ffmpeg flag ('-loop 1'), named scripts ('build-animatic.mjs', 'animatic/edl-v1.mjs'), and concrete prompt rules ('3-4 个 Then', 'Keep cartoon 句 + Do not realistic') — actionable enough that the absence of full code blocks is not penalized for an instruction-only skill.	3 / 3
Workflow Clarity	The 8-step 流程 is clearly sequenced with an explicit validation checkpoint ('Animatic 检查点（行为刹车）'), a '30 秒人工验收四问' checklist, roll judging via review-protocol, and a Common Mistakes 错误→后果→修复 table as an error-recovery feedback loop; batch/destructive generation is explicitly guarded, so it is not capped at 2.	3 / 3
Progressive Disclosure	References are well-signaled and one-level-deep ('→ 范例', '→ 范例+配方'), but no references/, scripts/, or assets/ bundle directories exist, so none of the cited files (prompt-book-v0.1.md, review-protocol-v0.1.md, build-animatic.mjs, etc.) can be verified as real shipped files, and several point to external repo paths — matching the 'references present but not backed by actual structure' level rather than the clean level-3 split.	2 / 3
	Total	10 / 12 Passed

Description

100%Weight 40%Scale 1-3

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A strong, well-structured description: concrete pipeline actions, natural trigger terms, explicit use-when/not-for boundaries, and clear disambiguation from sibling skills. Voice is third-person throughout, so no specificity penalty applies.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions and artifacts — 'charter+分镜+prompt book+素材账本+EDL 驱动的成片' plus the 图(关键帧)→视频(i2v)→后期(剪辑) pipeline — matching the 'lists multiple specific concrete actions' anchor.	3 / 3
Completeness	Explicitly answers what (the production line + Output artifacts) and when ('Use when:' clause) with explicit triggers, plus a 'Not for:' boundary; clearly the both-what-and-when anchor rather than the level below where 'when' is only implied.	3 / 3
Trigger Term Quality	'Use when: 做动画短剧/角色 IP 短片、猫咖日记系列新集、图生视频管线、剧情类 AI 视频' covers natural terms a user would actually say, matching the good-coverage anchor; not the level below because it supplies several common phrasings rather than a single generic keyword.	3 / 3
Distinctiveness Conflict Risk	'Not for: ...（用 video-forge）、...（用 image-generation）、...（用 ppt-forge）' carves a clear niche with distinct triggers and explicit redirects away from sibling skills, unlikely to trigger for the wrong skill.	3 / 3
	Total	12 / 12 Passed

Validation

87%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 14 / 16 Passed

Validation for skill structure

Criteria	Description	Result
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning
referenced_paths_exist	Referenced path issues: 1 missing	Warning

	Total	14 / 16 Passed

Repository: zts212653/clowder-ai
Path: cat-cafe-skills/anime-forge/SKILL.md
Commit: 80782c5

Reviewed: 11 days ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.