video-understand
Implement specialized video understanding capabilities using the z-ai-web-dev-sdk. Use this skill when the user needs to analyze video content, understand motion and temporal sequences, extract information from video frames, describe video scenes, or perform video-based AI analysis. Optimized for MP4, AVI, MOV, and other common video formats.
72
66%
Does it follow best practices?
Impact
Pending
No eval scenarios have been run
Advisory
Suggest reviewing before use
Optimize this skill with Tessl
npx tessl skill review --optimize ./skills/video-understand/SKILL.mdQuality
Discovery
89%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
This is a solid skill description that clearly communicates its purpose and when to use it. The explicit 'Use this skill when...' clause with multiple trigger scenarios is a strength, and the mention of specific video formats adds distinctiveness. The main weakness is that the capability descriptions could be more concrete—terms like 'analyze video content' and 'video-based AI analysis' are somewhat vague and could benefit from more specific action verbs.
Suggestions
Replace vague phrases like 'video-based AI analysis' and 'analyze video content' with more concrete actions such as 'detect objects across frames', 'generate scene-by-scene summaries', or 'track movement patterns'.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | It names the domain (video understanding) and mentions some actions like 'analyze video content', 'extract information from video frames', 'describe video scenes', but these are somewhat generic and not as concrete as listing truly specific operations (e.g., 'detect objects in frames', 'generate transcripts', 'track motion between frames'). | 2 / 3 |
Completeness | Clearly answers both 'what' (video understanding capabilities including analyzing content, extracting frame info, describing scenes) and 'when' (explicit 'Use this skill when...' clause with multiple trigger scenarios). The 'Use when' clause is explicit and detailed. | 3 / 3 |
Trigger Term Quality | Good coverage of natural terms users would say: 'video', 'analyze video', 'video frames', 'video scenes', 'MP4', 'AVI', 'MOV', 'motion', 'temporal sequences'. These are terms users would naturally use when requesting video analysis tasks. | 3 / 3 |
Distinctiveness Conflict Risk | The combination of video-specific triggers, the named SDK (z-ai-web-dev-sdk), specific video formats (MP4, AVI, MOV), and video-specific actions like 'temporal sequences' and 'video frames' creates a clear niche that is unlikely to conflict with image analysis, audio, or general file processing skills. | 3 / 3 |
Total | 11 / 12 Passed |
Implementation
42%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The skill provides highly actionable, executable code examples but is severely bloated by repeating the same API pattern dozens of times with only prompt text variations. The content would be far more effective as a concise overview showing the core pattern once, with a table of prompt templates for different use cases, and detailed integration examples split into separate reference files. The massive size actively harms usability by consuming context window budget.
Suggestions
Reduce to one core example showing the createVision pattern, then provide a concise table of prompt templates for different use cases (sports, education, moderation, etc.) instead of repeating the full function wrapper each time.
Move integration examples (Express.js, Next.js) and advanced use cases (batch processing, multi-turn conversation) to separate referenced files like INTEGRATIONS.md and ADVANCED.md.
Remove the overview bullet list of capabilities (Claude already understands what video analysis entails) and the 'Common Use Cases' numbered list which adds no actionable information.
Add explicit validation steps for batch processing workflows: verify URL accessibility before processing, validate response structure, and include a retry mechanism for failed videos.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Extremely verbose and repetitive. The same API pattern (createVision with video_url) is repeated 15+ times with only minor prompt variations. The overview lists capabilities Claude already understands, and many use-case functions (sports analysis, educational summarization, content moderation, quality assessment) are just prompt templates wrapped in identical boilerplate code. This could be reduced by 70%+ by showing the pattern once and listing prompt variations. | 1 / 3 |
Actionability | All code examples are fully executable, copy-paste ready JavaScript with proper imports, async/await patterns, and complete function signatures. CLI examples include specific flags and options. The Express.js and Next.js integration examples are production-ready with error handling and input validation. | 3 / 3 |
Workflow Clarity | The batch processing section includes error handling and rate limiting delays, which is good. However, there are no explicit validation checkpoints for video processing workflows (e.g., verify video URL is accessible before sending to API, validate response format). The recommended approach section mentions CDN upload and chunking but doesn't provide a clear step-by-step workflow for these operations. | 2 / 3 |
Progressive Disclosure | This is a monolithic wall of text with all content inline. Despite mentioning reference scripts in a scripts/ directory, all examples are fully embedded. The 10+ use-case functions (sports analysis, educational summarization, content moderation, etc.) should be in a separate reference file. The Express.js and Next.js integration examples could also be split out. No clear navigation structure beyond section headers. | 1 / 3 |
Total | 7 / 12 Passed |
Validation
90%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 10 / 11 Passed
Validation for skill structure
| Criteria | Description | Result |
|---|---|---|
skill_md_line_count | SKILL.md is long (917 lines); consider splitting into references/ and linking | Warning |
Total | 10 / 11 Passed | |
- Repository
- jjyaoao/HelloAgents
- Commit
07048a9
- Reviewed
Table of Contents
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.