Lanbow Ads Skills

Lanbow Ads Skills

End-to-end Meta Ads lifecycle management across 4 features: Strategy → Creative → Delivery → Review, forming a continuous optimization loop. Built by Lanbow — see the detailed guide for more information.

Feature Map

System Decision Tree

Starting a new campaign from scratch? → Start at Feature 1 (Strategy) → Feature 2 (Creative) → Feature 3 (Delivery)

Have strategy, need creatives? → Start at Feature 2 (Creative) → Feature 3 (Delivery)

Have creatives, need to launch? → Start at Feature 3 (Delivery)

Campaign already running, need to optimize? → Start at Feature 4 (Review) → loop back to Features 1-3 as needed

Quick ad hoc task? (list campaigns, check performance, upload image, etc.) → Use Feature 3 (Delivery / lanbow-ads CLI) directly

Full Lifecycle Loop

┌─────────────────────────────────────────────────────────┐
│                                                         │
│  Feature 1: Strategy Research                           │
│  ┌──────────────────────────────────┐                   │
│  │ Input: URL, objectives, audience │                   │
│  │ Output: insights report          │                   │
│  │   • key message                  │                   │
│  │   • key look                     │                   │
│  │   • ad angles + CTAs             │                   │
│  │   • audience segments            │                   │
│  └──────────┬───────────────────────┘                   │
│             │                                           │
│             ▼                                           │
│  Feature 2: Creative Generation                         │
│  ┌──────────────────────────────────┐                   │
│  │ Input: key message, key look,    │                   │
│  │        product info, audiences   │                   │
│  │ Output: ad images + prompts      │                   │
│  └──────────┬───────────────────────┘                   │
│             │                                           │
│             ▼                                           │
│  Feature 3: Ad Delivery                                 │
│  ┌──────────────────────────────────┐                   │
│  │ Input: creatives, targeting,     │                   │
│  │        budget, account           │                   │
│  │ Output: live campaign            │                   │
│  └──────────┬───────────────────────┘                   │
│             │                                           │
│             ▼                                           │
│  Feature 4: Post-Campaign Review                        │
│  ┌──────────────────────────────────┐                   │
│  │ Input: performance data          │                   │
│  │ Output: diagnosis + next actions │──── loop back ────┘
│  └──────────────────────────────────┘

Feature 1: Strategy Research

Skill: ads-strategy-researcher | Details: strategy-research.md

Perform pre-campaign market analysis using WebSearch + WebFetch. Produces a structured strategy report.

Trigger phrases: "ads strategy", "marketing plan", "competitor research", "market research"

Inputs: product URL, business objectives, campaign context, audience info, competitors

Key outputs that feed Feature 2:

Strategy types: Omni-Channel (Amazon + TikTok + Meta) or Meta-Only

Feature 2: Creative Generation

Skill: creative_gen | Details: creative-generation.md

Generate ad creative images via 2-step Gemini pipeline: proposals (gemini-2.5-flash) → images (gemini-3.1-flash-image-preview).

Trigger phrases: "generate ad", "create ad image", "ad creative", "generate creatives"

Requires: GEMINI_API_KEY (optional env var — only needed when using this feature; not required for Features 1, 3, or 4)

Inputs: product info, audience descriptions, requirements, CTA, aspect ratio, optional product image

Outputs: creative-proposals.json + ad-*.png images + reusable prompt templates

Connection to Feature 3: Upload generated images via lanbow-ads images upload, then reference image_hash in lanbow-ads creatives create.

Feature 3: Ad Delivery

Skill: lanbow-ads | Details: ad-delivery.md

Execute campaign setup and management via the lanbow-ads CLI.

Trigger phrases: "create campaign", "launch ads", "upload image", "show campaigns", "投放广告"

Requires: Meta Access Token, App ID, Ad Account ID (declared in requires.env). App Secret is optional — only needed for exchanging short-lived tokens to long-lived tokens via lanbow-ads auth exchange.

Credential setup (try in order):

1. Environment variables or platform secret fields (best): If META_ACCESS_TOKEN, META_APP_ID, and META_AD_ACCOUNT_ID are already set as environment variables (or via your platform's secret/credential fields), configure the CLI automatically:

   lanbow-ads config set --app-id "$META_APP_ID"
   lanbow-ads auth set-token "$META_ACCESS_TOKEN"
   lanbow-ads config set --account "$META_AD_ACCOUNT_ID"
   # Only if META_APP_SECRET is set (optional — needed only for token exchange):
   [ -n "$META_APP_SECRET" ] && lanbow-ads config set --app-secret "$META_APP_SECRET"

2. Ask the user to provide credentials directly (most common): If env vars are not set, tell the user exactly how to get each credential from Meta's web interface. Recommend the user use their platform's secret fields or environment variables rather than pasting credentials directly into chat. If direct input is the only option, only request the minimum credentials needed for the current task:

   • Access Token: Open https://developers.facebook.com/tools/explorer/ → select your App → click "Generate Access Token" → select permissions ads_management, ads_read, business_management → click "Submit" → copy the token
   • App ID: Go to https://developers.facebook.com/apps/ → select your App → App Settings → Basic
   • Ad Account ID: Go to https://adsmanager.facebook.com/ → find act_XXXXXXXXX in the URL or account dropdown
   • App Secret (only if needed for token exchange): App Settings → Basic → click "Show" next to App Secret

   Then run the CLI commands on their behalf. For a full walkthrough, see meta-account-setup.md.

3. lanbow-ads auth login (last resort): This opens a browser on the agent's machine for OAuth. It only works when you and the user are on the same machine. Do NOT attempt this by default. If the user can't open the auth URL or auth login fails, fall back to method 2 immediately — ask for credentials directly. Do NOT retry or keep sending OAuth URLs.

Credential scope limits: The agent must ONLY use provided credentials for lanbow-ads CLI commands. The agent must NOT log, echo, or store credentials in any file other than via the lanbow-ads config set and lanbow-ads auth set-token commands. The agent must NOT transmit credentials to any endpoint other than the Meta Marketing API (via the CLI).

Campaign creation sequence:

1. lanbow-ads campaigns create — create campaign (PAUSED)
2. lanbow-ads images upload / lanbow-ads videos upload — upload media
3. lanbow-ads creatives create — create creative with media
4. lanbow-ads adsets create — create ad set with targeting + budget
5. lanbow-ads ads create — create ad linking ad set + creative
6. Verify with lanbow-ads campaigns get / lanbow-ads ads get

Key rules:

• Budgets in cents (5000 = $50.00)
• Always create in PAUSED status, activate after review
• Use --json flag when output feeds into another step
• For OUTCOME_SALES: must include --promoted-object with pixel_id

Feature 4: Post-Campaign Review

Skill: post-campaign-review | Details: post-campaign-review.md

Analyze delivery data, diagnose performance issues, generate optimization recommendations.

Trigger phrases: "review campaign", "campaign diagnosis", "optimize ads", "复盘", "投后分析"

Data collection: Fetch via lanbow-ads insights get --json at multiple dimensions (campaign, ad set, ad, audience breakdowns, daily trends).

Diagnostic framework:

• Metric health check against benchmarks (CTR, CPC, CPM, frequency, ROAS)
• Root cause patterns: creative fatigue, audience saturation, targeting mismatch, budget issues, landing page problems

Output: Prioritized optimization plan (P0 immediate → P1 creative → P2 targeting → P3 strategic)

Feedback loop: Review findings feed back into:

• Feature 1 (new strategy if targeting mismatch or new opportunity found)
• Feature 2 (creative refresh if fatigue diagnosed)
• Feature 3 (budget/targeting adjustments, pause/activate ads)

Cross-Feature Data Flow

Strategy Report ──→ key message, key look, audiences ──→ Creative Gen
Creative Gen   ──→ ad images, proposals             ──→ Ad Delivery (upload + create)
Ad Delivery    ──→ live campaign + performance data  ──→ Post-Campaign Review
Review         ──→ optimization actions              ──→ Ad Delivery (adjust)
Review         ──→ creative refresh needs            ──→ Creative Gen (new round)
Review         ──→ strategy refinement needs         ──→ Strategy Research (re-analyze)

Security & Privacy