菜单
Activates when performing marketing inference on an ad image — extracting product/service, core message, target audience, brand codes, and mandatory copy from a visual. This is a shared analysis capability, not a user-facing workflow.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Activates during any ads-visual-fs command workflow when a structured user decision is needed. Provides the 4-section conversational format (Context, Explanation, Suggestion, Cost) and reusable patterns (Platform Selection, Visual Elements, Concept Selection, Next Action, Style Presets). This is a shared UX protocol, not a user-facing workflow.
Activates when generating creative concept variations at multiple conceptual levels (SAFE/BOLD/EXPERIMENTAL) for ad campaigns. Produces titles, rationales, and image generation prompts. This is a shared creative capability, not a user-facing workflow.
Activates during /create and /campaign workflows when the user needs to choose a visual style for ad creation. Provides 4 predefined styles (Warm Showcase, Bold Professional, Tech Forward, Aspirational Shadow) plus Custom. Each style defines required visual elements, composition, layout, and mood.
Activates after image generation to evaluate the output against brand compliance and brief alignment. Runs as step 5b in the standard workflow — between Generate and Review. Auto-retries on gross brand failures; advisory-only for fine-grained issues.
Activates during any ad creative work for Funding Societies — generation, refinement, resizing, or review. Enforces FS brand colors, typography, tone, logo usage, and regulatory compliance. This is a shared compliance capability that enhances all workflows.
Activates when analyzing a competitor ad for transferable creative strategies — extracting layout architecture, visual metaphor, color psychology, typography approach, imagery style, emotional appeal, messaging strategy, and attention mechanics, then mapping each to FS brand equivalents. This is a shared analysis capability, not a user-facing workflow.
| name | marketing-analysis |
| description | Activates when performing marketing inference on an ad image — extracting product/service, core message, target audience, brand codes, and mandatory copy from a visual. This is a shared analysis capability, not a user-facing workflow. |
| metadata | {"version":"0.7.1"} |
Structured marketing inference from ad visuals. Used by the Reimagine workflow.
This skill provides domain expertise when you need to analyze an ad image for marketing insights. You perform this analysis directly using your multimodal capabilities — do NOT call any MCP tool.
For each field, generate up to 5 candidate options ranked by confidence:
Present each insight as:
| Field | Inference | Confidence | Evidence |
|---|---|---|---|
| ... | ... | 0.85 | text overlay: "...", visual: ... |
Always flag uncertainties and assumptions separately.
After presenting analysis, always ask the user to confirm or edit insights before they are used downstream. This is critical — never proceed automatically.
