nsfw_detectionDetect whether an image contains NSFW (violent or sexually explicit) content and return an nsfw_score. 이미지가 NSFW(폭력적·선정적) 콘텐츠인지 탐지해 nsfw_score 를 반환합니다. detail=1 입력 시 세부 판정 결과를 함께 반환합니다. [호출당 10포인트]
| Name | Required | Description | Default |
|---|---|---|---|
| detail | No | 세부 판정 결과 포함 여부 (포함: 1, 미포함: 0, 기본값 0) | |
| image_url | Yes | 다운로드 가능한 https URL (허용 형식: image/jpeg, image/png, image/webp, image/bmp) (최대 50MB) |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already declare readOnlyHint=true, so the description need not restate that. It adds useful context about the detail parameter returning detailed results and the per-call cost of 10 points. These are beyond the annotations, but no additional behavioral traits like failure modes or rate limits are disclosed.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description contains redundant bilingual repetition: the same English and Korean sentence say the same thing. While overall length is short, the duplication wastes space and could be consolidated. The core purpose is front-loaded, but the structure is not tight.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
The description explains the output (nsfw_score) and the detail option, but does not clarify the score's scale, interpretation thresholds, or any failure behavior. Given there is no output schema, this leaves the agent uncertain about how to act on the result. Cost is mentioned, which helps, but overall completeness is only adequate.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 100%, so both `detail` and `image_url` are already documented (including formats and size limits). The description merely repeats the detail behavior ('detail=1 입력 시 세부 판정 결과를 함께 반환') without adding new meaning beyond the schema, so it sits at the baseline.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool detects NSFW content in an image and returns a score, using a specific verb and resource. It distinguishes from siblings like face_detection and image_similarity by the domain, but does not explicitly name them or contrast, so it misses the top mark.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
No guidance is provided on when to use this tool versus alternatives. It does not mention scenarios like content moderation or say to prefer image_similarity for similarity tasks, leaving the selection burden entirely on the agent.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Add one secure layer between your agents and this server.
Each tool has a clearly distinct purpose: video thumbnail extraction, face detection, image similarity, NSFW detection, video-to-audio conversion, and word cloud generation. There is no overlap between any two tools, so agents can easily select the correct one.
All tool names use snake_case, but the pattern is mixed: most are noun-based (face_detection, image_similarity, nsfw_detection, word_cloud) while extract_video_thumbnail starts with a verb and video_to_mp3 is a conversion style. This inconsistency makes the naming slightly less predictable.
With 6 tools, the count is well within the typical 3-15 range and feels well-scoped for a media processing toolkit. Each tool adds a distinct capability without being redundant or overwhelming.
The server covers a variety of media operations (video, image, audio, text) but lacks common vision tasks like OCR, object detection, or image classification that one might expect from a 'Vision' server. The set feels like a collection of standalone utilities rather than a complete domain coverage, though no critical gaps cause agents to fail.