infinitetalk

音频驱动的稀疏帧视频配音工具，支持音频驱动的 Video-to-Video 和 Image-to-Video 生成，实现精准的唇形、头部、身体姿态同步，支持无限时长视频生成

Quality

76%

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

Fix and improve this skill with Tessl

tessl review fix ./skills/infinitetalk/SKILL.md

Quality

Content

85%

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

The body is highly actionable with executable commands, clear per-mode workflows with validation, and well-structured one-level-deep references to verified bundle files. The main weakness is minor redundancy between the capability list, mode descriptions, and notes sections.

Suggestions

De-duplicate the hardware/quantization guidance so it appears in only one section (前置准备 or 注意事项), referencing the other rather than restating.

Collapse the repeated 能力包含 bullets that overlap with the per-mode 操作步骤 to reduce token cost.

Consider moving the parameter reference table into the usage_examples reference file to keep the body a tighter overview.

Dimension	Reasoning	Score
Conciseness	The body is mostly lean with parameter tables and copy-paste commands, but it repeats content across sections (capability bullets restated in mode list, hardware/quantization advice duplicated in 注意事项) that could be tightened.	2 / 3
Actionability	Provides fully executable bash commands with real flags, a concrete parameter reference, and an existing inference script — copy-paste ready with specific examples.	3 / 3
Workflow Clarity	Each mode has a clear numbered sequence with an explicit 验证输出 (validate output) checkpoint and an error-recovery section for out-of-memory failures, giving feedback loops for fragile operations.	3 / 3
Progressive Disclosure	A concise overview links one level deep to real, verified reference files (model_download, environment_setup, usage_examples) and the inference script, with a 资源索引 index for easy navigation.	3 / 3
	Total	11 / 12 Passed

Description

67%

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

The description is specific and distinguishes the skill well, but it omits an explicit 'when to use' trigger clause and underuses natural keyword variations that a user might say. Adding a Use-when clause with common terms like 数字人/虚拟主播 would lift completeness and trigger quality.

Suggestions

Append an explicit trigger clause, e.g. 'Use when generating audio-driven digital-human videos, video dubbing, or virtual-anchor content.'

Broaden trigger keyword coverage to include user-natural terms like 数字人, 虚拟主播, talking-head, and 口型同步.

Remove the repeated 音频驱动的 phrasing to tighten the first clause and avoid redundancy.

Dimension	Reasoning	Score
Specificity	Names multiple concrete capabilities — Video-to-Video, Image-to-Video, lip/head/body posture sync, and unlimited-duration generation — rather than vague abstractions.	3 / 3
Completeness	Clearly states what the skill does, but provides no explicit 'Use when...' trigger guidance; per the guidelines a missing trigger clause caps completeness at 2.	2 / 3
Trigger Term Quality	Has relevant natural terms like 视频配音 and 音频驱动, but lacks keyword variation (e.g. 数字人/虚拟主播 appear only in the body) and offers no explicit trigger phrasing.	2 / 3
Distinctiveness Conflict Risk	Targets a clear niche (audio-driven talking-head video generation) with distinct triggers, making overlap with generic video or image skills unlikely.	3 / 3
	Total	10 / 12 Passed

Validation

93%

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 — 15 / 16 Passed

Validation for skill structure

Criteria	Description	Result
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	15 / 16 Passed

Repository: anbeime/skill
Path: skills/infinitetalk/SKILL.md
Commit: f595e96

Reviewed: about 12 hours 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.