Hubbestaiskill.com

Home / All Skills / Design & Motion / vibeshot-candid-photography

Design & Motion · vibeshotclub/vsc-skills

vibeshot-candid-photography

Generates portrait photography prompts featuring authentic everyday life, candid snap aesthetic, unconventional camera angles, and pronounced spatial layering. Ideal for requests involving candid photography, lifestyle portraits, Korean/Instagram-style portraits, natural foreground occlusion, unconventional composition, or batch photography prompt generation.

L1 Basic · Source registration onlyV1 Commit-PinnedDual-pass Audit

Author vibeshotclubGitHub vibeshotclub/vsc-skills ↗Stars 832License 仓库未提供 LICENSE 文件（find 全仓无 LICENSE*；GitHub API license 字段为 null）Commit fcee81beb6

Control levels describe skill capabilities and suggested control measures; they do not represent quality or safety certifications, nor do they dictate usability. Usage decisions rest entirely with users and administrators. Different agent hosts apply varying constraints; do not assume hosts provide automatic protection. Third-party skill copyrights belong to their original authors. Use at your own risk.

1How It Works & Mechanism

Implements a structured checklist approach for the 'random variable pool': dozens of candidates across twelve dimensions (expression, outfit, scene, candid moment, shot scale, focal length, camera angle, composition, foreground, lighting, color, photographic state), naturally fused into a prompt through model sampling.

vibeshot-candid-photography/SKILL.md

每组提示词从以下变量池中随机组合。

The vast majority of the 319 lines in SKILL.md consists of these pool entries (e.g., 50+ scene options, 25+ camera angle options).

Anchors a fixed subject profile by default, unless explicitly overridden by the user.

vibeshot-candid-photography/SKILL.md

一个极具镜头感的韩国 INS 网红，皮肤白皙，傲人身材。

Explicitly stated immediately after: 'Unless the user explicitly modifies the subject, always maintain this subject setting.'

Treats batch processing as a first-class citizen: defaults to 10 sets, each ready for direct copy-pasting and distinctly varied from one another.

vibeshot-candid-photography/SKILL.md

默认输出 10 组完整提示词。

The 'Batch Diversity Rules' section strictly prohibits consecutive similar scenes or camera perspectives within the same batch.

Prevents repetition via a prioritized blacklist of duplication dimensions: scene, candid moment, shot scale, focal length, camera angle, composition, foreground, and lighting.

vibeshot-candid-photography/SKILL.md

场景、瞬间状态、景别、焦段、机位、构图、前景、光线。

The same section mandates actively mixing close-ups/environmental shots, low/high angles, wide-angle/telephoto, indoor/outdoor, and quiet/dynamic moments.

Strictly restricts output formatting: outputs only numbered full prompts, forbidding structured key-value fields and procedural explanations.

vibeshot-candid-photography/SKILL.md

不要输出参数解析、变量说明、创作思路或组合过程。

Explicitly bans field-style notations such as 'Scene:/Outfit:/Focal Length:/Angle:'.

Grants user-specified conditions top priority; locked dimensions are excluded from random sampling.

vibeshot-candid-photography/SKILL.md

用户指定的任何条件优先级最高。

For example, if 'lotus pond + BM style + 28mm + ultra-low angle' is provided, these four are locked while remaining dimensions continue to be randomized.

Sampling is not a mechanical fill-in-the-blank process: pools only guide direction, and the model reassembles elements using natural photographic language with reasonable creative freedom.

vibeshot-candid-photography/SKILL.md

变量池负责提供方向，不要求将抽中的每一个词原样写进最终提示词。

The prompt generation section outlines the assembly sequence: subject → vibe/expression → outfit → scene → candid moment → shot scale/focal length → camera angle → composition → foreground → lighting → color → photographic state → global requirements.

Anchors aesthetic direction using a strict 'Avoid' list: anti-studio, anti-commercial backdrop, anti-plastic skin, anti-synthetic AI mannequin look, and strictly forbids straightening out unconventional compositions for superficial neatness.

vibeshot-candid-photography/SKILL.md

不要为了让人物“漂亮”而自动把非常规构图修正成普通人像。

The same section specifies 'do not have all subjects look directly at the camera,' 'do not center all subjects,' and 'avoid excessive skin smoothing.'

Translates the candid feel into actionable movement requirements: prioritizes in-motion action states over completed standard poses.

vibeshot-candid-photography/SKILL.md

优先选择动作正在发生中的瞬间，而不是动作已经完成后的标准姿势。

The foreground pool also requires 'foreground must naturally intrude into the frame and create occlusion' and 'do not remove occlusion simply to show the subject completely.'

2Core Capabilities

01Twelve-dimension random variable pool (expression, outfit, scene, moment, shot scale, focal length, angle, composition, foreground, lighting, color, photographic state)

02Batch output defaulting to 10 sets, with support for user-specified counts

03Constraint locking: user-provided scenes, outfits, focal lengths, angles, etc., are excluded from randomization

04Color palette constraint to avoid visual clutter (maximum 3–4 primary color blocks per image)

05Unconventional angles and extreme compositions as default aesthetics (ground-level, low-angle tilt-up, over-the-shoulder, frame-within-frame, edge-aligned subjects)

06Anti-pattern checklist targeting indoor studio photography to achieve authentic everyday life textures

07'Prompt-only' delivery mode: generates zero extraneous content outside the prompts

08Fixed numbering format (### 01 / ### 02) with no post-generation photographic analysis

4Recommended Governance & Evidence

Suggested Controls for Safe Operation

For Agent Runtime

Can be auto-loaded

For Administrators

Register source and version

Evidence & Triggers (File Paths & Excerpts)

SKILL.md: Pure prompt directives and rule text without script execution or external impact

Audit Notes: Contains no scripts, network transmissions, credential reads, or local writes; the directory consists solely of SKILL.md, README.md, and agents/openai.yaml. Points of attention include content inclinations (defaults to an influencer portrait profile with body descriptions, scenes like pools/lotus ponds) and potential token usage and platform compliance considerations stemming from the default 10-batch output.

5Second-Pass Verification

[ok] 变量池维度清点（防止把「十二维」写成凑数） — SKILL.md 小标题实测为：表情随机、服装随机、场景随机、瞬间状态随机、景别随机、镜头焦段随机、机位随机、构图随机、前景随机、光线随机、色彩随机但克制、摄影状态随机 —— 共 12 个，表述与文中一致。
[ok] 输出规则与 README 示例是否一致 — README 示例输出为 ### 01/### 02 编号 + 整段自然语言，无字段式排版、无变量解释，与 SKILL.md「不解释随机到了什么变量」等要求一致。
[ok] 外部资源/依赖条目（本 skill 无） — SKILL.md/README 均未要求任何 CLI、API、包或网络；全仓 http 扫描本目录零命中。
[ok] 安全结论反例搜索 — 该目录三个文件全为明文 Markdown/YAML；无脚本片段、无 base64、无网络或写盘调用。
[ok] 默认主体是否为可覆盖的默认值（而非硬绑定） — SKILL.md 明确『除非用户明确修改主体，否则始终保持这一主体设定』，README 也给出「主体改为短发中国女摄影师」的示例输出，默认值可被用户条件覆盖。
[ok] 与相邻技能的路由边界 — vsc-distinction 写明『无指定角色或夏季泳装 POV 要求时适用。仅出现泳池或夏天不等于泳装 POV』，与 tests/vsc-routing-cases.md 中「普通成年人物在泳池边喝咖啡，穿日常衣服」应路由到本技能的用例一致。

6Verdict & Scenarios

Direct mechanism for batch diversity: 12-dimensional variable pool + explicit anti-repetition rules minimize overlap across 10-prompt batches.

Strict aesthetic enforcement (anti-studio, anti-plastic skin, anti-centered posing, preserving unconventional framing) drastically reduces artificial 'AI mannequin' look.

Clean delivery discipline: outputs only copy-paste-ready prompt paragraphs without structured fields or explanations, ideal for batch logging and automation.

User preferences take precedence, facilitating controlled experiments (e.g., locking lotus pond + BM style + 28mm + low angle while randomizing the rest).

Best for: Ideal for batch-generating portrait prompts demanding authentic candid lifestyle vibes, unconventional camera angles, and natural foreground framing: Korean/Japanese everyday lifestyle portraits, social media stock visuals, or scenarios requiring a batch of stylistically coherent yet visually distinct prompts; also well-suited for controlled prompt experiments by locking specific constraints (scene, outfit, focal length, angle).

Not for: Not suitable for specific character cosplay or anime/2D candid shots (use character-candid-photography instead); not suitable for summer swimwear or partner/friend-POV single photo requests (use summer-boyfriend-pov, which explicitly notes that swimming pools or summer settings alone do not constitute swimwear POV); not suitable for scenarios requiring direct image generation; not suitable for experiments requiring deterministic, exactly reproducible sampling (no seed mechanism).

Install Agent-Ready

① Verified Mirror 2026-09-15

Option A · Direct DownloadDownload vibeshot-candid-photography.tar.gz

sha256: aa0ba7e90b693ca9…

Option B · Copy Guide for Agent

Agents automatically download and verify mirrors from this JSON guide.

② Upstream GitHub

GitHub Repo ↗

Pinned to commit fcee81beb6.