timesmkt3

Conducts structured marketing intelligence research for a given business niche or topic using the local Tavily AI SDK. Use when the user asks to "research a market", "analyze competitors", "generate campaign insights", "find marketing angles", "research content ideas", "run market research", or "build a research brief". Outputs machine-readable structured JSON, a Markdown report with mermaid graphs, and an interactive HTML report with Chart.js and brand styling. These outputs are ready for downstream agents (video, image ads, copywriter) to consume directly. Always use this skill when the pipeline needs a strategic research layer before generating creative assets.

Generates structured static ad creative layouts for social media advertising as design JSON. Use when the user asks to create an image ad, design a static ad, generate an ad creative, produce a layout for Instagram, make a banner ad, or says "create an ad", "design a social ad", "generate ad layout", "make an image creative", or "build a static ad for Cold Brew". Outputs design JSON rendered by the frontend — not raw images. Always use this skill when static ad creative generation or layout JSON is needed for Cold Brew Coffee Co.

Handles media hosting, publishing readiness, and post scheduling for campaign outputs. Use when the user asks to "distribute content", "upload media", "schedule posts", "publish to Instagram", "publish to YouTube", "prepare for publishing", "generate a publish file", or "host campaign assets". Uploads output files to Supabase, generates public media URLs, compiles platform metadata from Research and Copywriting Agent outputs, and produces a Markdown publish advisory file per campaign. CRITICAL: Never execute actual posting unless the user explicitly references the generated "Publish <task_name> <date>.md" file by name. Always use this skill when the pipeline is in the distribution phase.

Generates platform-specific marketing copy for Threads, Instagram, and YouTube using structured research output from the Marketing Research Agent. Use when the user asks to "write copy", "generate captions", "create social posts", "write a YouTube description", "generate Instagram copy", "write Threads posts", or "create marketing copy". Consumes research JSON from the outputs/ folder and brand guidelines from knowledge/ to produce consistent, platform-native copy across all three channels. Always use this skill after the marketing-research-agent has run, or whenever platform copy needs to be generated for a campaign.

Runs the full AI content pipeline as a single coordinated workflow using BullMQ and Redis. Use when the user asks to "run the pipeline", "start the campaign", "kick off the full workflow", "orchestrate the agents", "run all agents", or provides a job payload with task_name and task_date. Manages agent dependency ordering, validates skip conditions, enqueues BullMQ jobs, and tracks completion status. Always use this skill when multiple agents need to run in sequence or parallel for a single campaign. Do NOT use for running a single agent in isolation.