360 Secure SkillHub
360 安全技能中心

Ads Execution Hub

面向媒体投放团队提供跨 Meta、Google Ads、TikTok Ads、YouTube Ads、Amazon Ads、Shopify Ads 及程序化 DSP 的广告操作支持:规范化目标与 KPI、给出竞价与预算分配建议、设定测试与放量规则、输出监控告警触发条件与操作交接清单,并在数据缺失或账户受限时给出降级与升级路径。

danyangliu-sandwichlab
danyangliu-sandwichlab
热度 53.42
最近更新 2026-05-11 17:30:24
分类:营销依赖:纯提示词应用场景:市场分析 / 工作流 / 模板 / 报告撰写

安装技能

复制下面的话术发给你的 Agent,它会通过 secure-skills 检索并安装该技能。

执行命令安装此 skill: npx secure-skills-cli add ads-execution-hub
Metadata
name
ads-execution-hub
description
Ads Execution Hub control skill for ad campaign management and optimization across Meta (Facebook/Instagram), Google Ads, TikTok Ads, YouTube Ads, Amazon Ads, Shopify Ads, and DSP/programmatic.

Ads Execution Hub

Purpose

Core mission:

  • Serve as the dedicated ad operations and optimization interface.
  • Manage planning, launch, monitoring, and scaling across ad channels.
  • Standardize decision policies for bidding, budget, and performance recovery.
  • Output clear operator actions for media teams.

When To Trigger

Use this skill when the user asks for:

  • campaign setup, optimization, or scaling in one or more channels
  • budget and bidding decision support with performance constraints
  • anomaly diagnosis and recovery actions for live campaigns
  • cross-channel media operation playbooks

High-signal keywords:

  • ads execution hub, run ads, campaign, media buyer
  • bidding, budget, allocation, optimize, scale
  • cpa, roas, performance, monitor, abtest

Input Contract

Required:

  • campaign_objective
  • channel_scope
  • budget_constraints
  • recent_performance_snapshot

Optional:

  • creative_state
  • audience_state
  • tracking_health
  • policy_or_account_flags

Output Contract

  1. Campaign Action Plan
  2. Bidding and Budget Policy
  3. AB Test and Scale Model
  4. Monitoring and Alert Plan
  5. Operator Handoff Checklist

Workflow

  1. Normalize objective and KPI constraints.
  2. Evaluate channel readiness and structure quality.
  3. Produce bid and allocation actions.
  4. Attach testing and scaling rules.
  5. Return monitoring triggers and operator checklist.

Decision Rules

  • If measurement confidence is low, limit scale and improve tracking first.
  • If ROAS is stable above threshold, allow staged budget increases.
  • If CPA is unstable, reduce concurrency of experiments.
  • If anomaly risk is high, prefer containment actions first.

Platform Notes

Primary scope:

  • Meta (Facebook/Instagram), Google Ads, TikTok Ads, YouTube Ads, Amazon Ads, Shopify Ads, DSP/programmatic

Platform behavior guidance:

  • Keep channel recommendations execution-specific and auditable.
  • Align bid logic with each platform's optimization mechanics.

Constraints And Guardrails

  • No irreversible changes without rollback conditions.
  • Keep every recommendation tied to KPI impact.
  • Respect policy and account health constraints.

Failure Handling And Escalation

  • If required platform data is missing, return minimum data request list.
  • If policy or account block appears, route to compliance/account helper.
  • If spend risk is severe, trigger emergency control mode.

Code Examples

Campaign Control Spec

objective: improve_roas
channels: [Meta, GoogleAds, TikTokAds]
budget_mode: staged_scale
cpa_ceiling: 42
roas_floor: 2.5

Alert Trigger Rule

if roas_drop_pct > 20 and spend_up_pct > 25:
  severity: high
  action: cap_budget_and_notify

Examples

Example 1: Launch and stabilize

Input:

  • New campaign across Meta and TikTok Ads

Output focus:

  • launch checklist
  • first-week controls
  • fallback rules

Example 2: Scale after validation

Input:

  • Stable ROAS for 10 days

Output focus:

  • scale ladder
  • bid policy updates
  • monitoring checkpoints

Example 3: Cross-channel anomaly

Input:

  • Spend surge, mixed conversion signals

Output focus:

  • anomaly triage
  • containment actions
  • next validation steps

Quality Checklist

  • Required sections are complete and non-empty
  • Trigger keywords include at least 3 registry terms
  • Input and output contracts are operationally testable
  • Workflow and decision rules are capability-specific
  • Platform references are explicit and concrete
  • At least 3 practical examples are included