Nano Banana Kling Ad Workflow

Recreate low-budget AI video ad workflows using Nano Banana image generation plus Kling 3.0 video synthesis with dialogue, including prompt design, scene planning, cost control, and export handoff. Use when a user wants to produce a cinematic ad quickly (often in a few hours) with a small credit budget, or asks for a Deon-style Nano Banana + Kling pipeline.

Security Scan

This skill is an instruction-only checklist for creating ads with Nano Banana and Kling 3.0; its requirements and instructions are consistent with that purpose and it does not request unexplained credentials or install code.

Purpose & Capability

The name/description match the instructions: a two-stage Nano Banana (image) → Kling 3.0 (video) ad pipeline. No unrelated binaries, env vars, or config paths are requested, so the declared capabilities align with what the skill actually asks the agent to do.

Instruction Scope

SKILL.md stays within the ad-production workflow: shot planning, prompt patterns, generation/animation steps, cost-tracking and deliverables. It does not instruct reading system files, other skills' configs, or exporting data to unexpected endpoints. The only minor openness: it says 'If missing, ask for only the minimum required details and proceed,' which grants some discretion but is reasonable for a production checklist.

Install Mechanism

No install specification or code files are present (instruction-only). There is nothing written to disk or fetched at install time, which minimizes supply-chain risk.

Credentials

The skill does not require environment variables, API keys, or credentials in its metadata. That is proportionate: the workflow describes third-party services (Nano Banana, Kling) but does not ask for unrelated secrets. Note: actual runtime use with those services would typically require API keys, but those are not requested by this skill itself.

Persistence & Privilege

always is false and the skill is user-invocable; it does not request persistent or elevated privileges nor does it modify other skills or system-wide settings.

Assessment

This is an instruction-only workflow checklist and is internally coherent. Because the skill's source and homepage are unknown, consider the following before installing or using it: (1) at runtime you or your agent will likely need Nano Banana and Kling API credentials—provide those only if you trust the integration and understand costs; (2) confirm any provider Terms of Service, credit costs, and content/licensing implications (especially for likeness/voice use); (3) be cautious about feeding sensitive data into generation prompts; and (4) if you want higher assurance, ask the publisher for provenance (homepage, author) or prefer a skill with a known source. Overall risk is low for this instruction-only checklist, but verify external service credentials and costs when you actually run the pipeline.

Overview

Build a short ad from scratch using a fast two-stage pipeline: generate stills in Nano Banana, animate them in Kling 3.0, then stitch a publishable cut. Optimize for speed, visual consistency, and low spend.

Workflow

1) Define outcome before generating assets

Capture these constraints first:

• Product or story concept
• Audience and tone
• Target duration (15s, 30s, or 45s)
• Delivery format (X, TikTok, Reels, YouTube)
• Budget ceiling in credits

If missing, ask for only the minimum required details and proceed.

2) Build a shot list

Create 5-9 shots with:

• Shot number
• Scene goal
• Subject + environment
• Camera style
• On-screen line or dialogue intent

Keep each shot prompt short and concrete.

3) Generate base visuals in Nano Banana

For each shot:

• Prompt for one clear hero frame
• Keep recurring anchors stable (character traits, wardrobe, color palette, lens style)
• Generate 2-4 variations max, pick one

If consistency drifts, add explicit anchor text to the next prompt.

4) Animate in Kling 3.0

Import selected stills into Kling 3.0 and add:

• Motion direction (camera push, pan, dolly, parallax)
• Dialogue or narration intent
• Timing per clip (usually 2-5s)

Prefer subtle motion over aggressive movement unless the concept requires action-heavy pacing.

5) Assemble final cut

Sequence clips by narrative flow:

• Hook (first 1-2 shots)
• Value demonstration
• Clear CTA

Add captions if platform autoplay is likely muted.

6) Track cost and output quality

After generation, report:

• Total clips generated
• Credits used and estimated cost
• Final runtime
• Export ratio(s)
• What to improve in next iteration

Prompt pattern

Use this compact prompt shape for each Nano Banana shot:

"[subject], [action], in [environment], [lighting], [camera framing], [style anchors], ultra-clean composition, ad-grade, no text overlays"

Use this compact Kling prompt shape:

"Animate this still with [motion], keep subject identity stable, cinematic realism, [timing], [dialogue/emotion cue], smooth transitions"

Fast defaults

• Runtime target: 20-30 seconds
• Shot count: 6
• Variations per shot: 3
• Clip length: 3-4 seconds
• Revision passes: 1 content pass + 1 polish pass

Failure handling

• If faces drift: repeat identity anchors and reduce motion complexity
• If scenes look noisy: simplify prompts and reduce style stacking
• If cost rises too fast: reduce variations and shorten shot list
• If timeline slips: ship a 15-second cut first, then extend

Deliverable format

When executing this skill, output:

1. Final shot list
2. Prompt set used
3. Generation and edit log
4. Final export summary
5. Next-iteration recommendations